Labour market transitions and risks of exclusion.

Abstract

Transitional labour markets (TLMs) are used as a descriptive device to investigate transitions between employment and other activities in Australia, and to identify groups most at risk of labour market exclusion. Using data from the Household, Income and Labour Dynamics of Australia (HILDA) survey, this article examines transitions between full-time work, three categories of part-time work, unemployment, unpaid caring, study, retirement, and 'not in the labour force but not studying or caring'. The results show that around 20 per cent of the adult population each year makes a transition into, out of, or within employment, combined with other activities such as study or caring. However, some groups are at much greater risk of 'exclusionary transitions'. These include workers who were retrenched within the last decade, those with less than year 12 education, people with long term health problems, migrants with poor English, parents with children under 5 years of age (especially sole parents), women and Indigenous people. Policy directions consistent with a TLM approach to improved employment participation for excluded groups include: a youth guarantee, retraining for workers and linked employment and training programs for the long-term unemployed.

Introduction: the Policy Challenge

Social policy-makers face a major challenge following the breakdown of the 'wage earners' welfare state', which was premised on full-time male employment and female household production, tariff protection, centralised wage fixing and full employment as a central policy goal (Castles 1996).

Enterprise bargaining has largely replaced centralised bargaining, making wage rises conditional upon productivity improvements and detaching workers with less market power from the gains made by stronger unions (Buchanan and Watson 2000). The industrial relations changes flagged by the Howard government in 2005 for implementation in 2006 will exacerbate this trend. Economic policy is primarily concerned with controlling inflation, improving competitiveness and fostering economic growth with the assumption that increased employment will occur as a result. Tariff protection has been replaced by an emphasis on open borders to promote international trade.

A common theme has been a shift from protection through risk prevention (tariff protection, centralised bargaining, the breadwinner/family wage) to one of flexibility and adaptability. This was partly driven by neo-classical economic doctrines and Third Way variants (e.g. Giddens 1998), although policy makers during the 1980s were also influenced by the examples of European social democratic regimes (such as Sweden and Norway) which had adopted an international trading economic policy with accompanying restructuring tied to significant investment in active labour market policies (Cass 1988).

In Australia, economic growth since 1960 has been strongly driven by rising levels of education (Chou 2003), and demand for skilled workers is likely to continue (Keating 2003). Growth over the medium term seems contingent on developing key 'knowledge industries' such as biotechnology, information and communications and technology and health, education and business services. This requires staff who are more highly educated and skilled, and who are willing to learn new skills in order to adapt.

At the same time, service industries have generated many low-skilled and low-paid jobs, often part-time and casual. The IT revolution has also created low-skilled low-paid positions (such as call centre workers), so it cannot be assumed that the knowledge economy will create only highly paid and skilled jobs. Overall, while many lower paid part-time jobs have been created, if employment is measured in terms of total hours of work, the greatest demand has been in higher skilled occupations (Keating 2003).

Despite record economic growth, Australia is becoming more polarised in terms of income and wealth and unemployment and underemployment remain high. Under-utilised labour--the unemployed and underemployed and those marginally attached to the workforce--is approximately double the unemployment rates (Australian Bureau of Statistics 2004), and recent analyses put the rate of 'labour wastage' at 15 per cent of the labour force (Burgess, Mitchell and Preston 2003).

The rise of non-standard employment has produced both gains and losses. Many people, for example, prefer part-time employment, particularly students and parents. However, for some workers, we also see new forms of insecurity and risk: irregular or unpredictable income, uncertain working time and hours, poor access to training, reduced opportunities for participation in decision making, vulnerability to OHS hazards, job insecurity and low pay (Macdonald and Holm 2001; Standing 2002).

One consequence of sustained unemployment and the growth of two-job and jobless households has been a shift in the means of income redistribution. Intra-family private redistribution of income from working husbands to non-working wives is being replaced by public transfers from working to non-working households, leading to growing expenditure on social security payments and parallel reforms aimed at cutting welfare spending.

Since the mid-1990s, a range of changes to unemployment payments have been introduced; these include elements of workfare (Work for the Dole) first for young people and then for most others on unemployment benefits, cuts to wage subsidy and training programs, and more onerous job search requirements. Most of these changes have emphasised a 'work-first' approach to employment assistance rather than a skills development or human capital approach.

Flexibility and Risk

The trends outlined above reflect the conclusions of Esping-Andersen's work (published in 1999) on the challenges facing various welfare state regimes. As traditional forms of protection have been dismantled and Australia has progressed further towards a pure liberal model, risks have been increasingly individualised. The economy has generated both demand for high skilled workers and numerous low-skill service jobs; and, while increased flexibility has led to jobs growth, it has also allowed underemployment and precarious work to flourish. Since the protection enshrined in the labour market has been wound back, more people have had to rely on the social security system and are forced to live on very low incomes; sometimes complemented with part-time or casual work taxed at very high marginal rates.

Previous Australian research has argued that these risks require new forms of protection, which will involve greater adaptability and flexibility combined with a concern for better support and redistribution towards those at greatest risk (Ziguras et al. 2004). This reflects a similar debate in Europe on the flexibility-security nexus (Muffels, Wilthagen and van den Heuvel 2002).

Transitional Labour Markets

Schmid's (1998, 2000, 2002) notion of transitional labour markets (TLMs) has been proposed as one way of responding to the changes described above (Ziguras et al. 2004). TLMs describe periods during which people move between full-time work and other activities such as caring, education and retirement. Schmid uses TLMs to propose policies which combine the flexibility and adaptation which are the hallmarks of the active labour market approach, with an aspiration for a fairer distribution of paid employment. In effect, Schmid proposes that active labour market policies be expanded to allow people to move both in and out of paid employment more easily.

TLMs require an 'active approach' in that they should both provide not only financial support but other assistance (such as for training and caring for others) for people to make transitions between employment roles. They should embody the principle of increased capacity building central to active labour market policy, but are also directed to people who wish to move out of work, as well as those who are unemployed. TLMs should aim to:

* Combine employment and other useful activities not valued on the market

* Provide a combination of wages, transfer payments and other income sources

* Ensure an agreed entitlement 

* Finance employment and capacity-building activities instead of unemployment.

Schmid (1998. p. 10) suggests that five major life-course transitions can be identified:

* Between education and employment

* Between caring and employment

* Between unemployment and employment

* Between employment and retirement

* Between precarious and permanent employment.

The TLM framework can be thought of as both normative and descriptive; that is, it can be used to describe the current operation of the labour market and also to design policies to change the way the labour market is structured.

Research Aims

This study investigates the extent to which labour market transitions of the type described by Schmid (1998, 2000, 2002) exist in Australia, to document the characteristics of those most at risk of labour market exclusion (or 'exclusionary transitions'), and to suggest some policy directions to provide greater protection for those most at risk. This article is divided into three main sections; description of current transitions, analysis of exclusionary transitions and future policy.

The first section, explores whether TLMs provide a useful lens for viewing the Australian labour market, by investigating the first four of the transitions listed above. Since part-time work has grown dramatically over the last decade, transitions between full-time employment, part-time work combined with other activities (focussing on caring and education), unemployment, and other activities outside the labour market (caring, education and retirement) are examined.

The second section of the paper investigates the risk factors for exclusion. Schmid (1998) argues that TLMs can also be used as a normative device to describe the quality of transitions. Cebrian, Lallement and O'Reilly (2000) suggest it may be possible to distinguish three types of transitions; integrative, maintenance and exclusionary. Integrative transitions would allow people outside paid employment, such as the unemployed, students and carers, to move into full-time employment via part-time work. Maintenance transitions would allow people in employment to maintain 'employment continuity by moving between different working time regimes' (Cebrian, Lallement and O'Reilly 2000, p. 4). Finally, exclusionary transitions would be represented by periods of part-time or temporary work within a longer pattern of unemployment or non-employment. This would effectively mean transitions only between the peripheral labour market and non-work.

Muffels, Wilthagen and van den Heuvel (2002) defined total exclusion as working for no period over three years, and partial exclusion as working less than 50 per cent of the time. For reasons described below, exclusionary transitions are defined in this study as having spent less than 50 per cent of the year in employment. For this part of the analysis, only those in the labour market or who are marginally attached to the labour market are included. The research explores which people are at risk of poor transitions by examining risk factors for flows into unemployment and exclusion from paid work.

Identifying Transitions Using HILDA Data

This analysis used data from the Household, Income and Labour Dynamics of Australia (HILDA). HILDA is a longitudinal study of a stratified random sample of the Australian population. The data used in this analysis was collected in the first two waves in 2001 and 2002. In Wave 1, 2001, a total of 13,969 adults over the age of 15 were included. In Wave 2, 2002, 85.9 per cent of these were included--an attrition rate of 14.1 per cent, which is fairly low by international comparison (Watson and Wooden 2002). This leaves data on 11,993 people available for this analysis.

Schmid's framework was used to classify respondents into nine situations according to their employment status and other activities: full-time work, three categories of part-time work, unemployed, unpaid caring, study, retired, and not in the labour force but not studying or caring. Since part-time work has grown significantly, and appears to be a key factor in transitions, I distinguished between people studying or caring with part-time work from those in these categories not in the labour force.

In the case of part-time work, three categories were used, based on the main reason people stated they were working part-time. 'Part-time work (care)' indicates they were caring for a child under 16 years of age, and 'part-time work (study)' means they were undertaking post-secondary education. 'Part time work (other)' includes people working part time because of a disability, who were seeking full-time work, who preferred part-time work, or were undertaking part-time work for other reasons not stated. The last category--'not in the labour force (other)' includes those people who were not in the labour force according to conventional definitions, were not caring or studying and who had not retired permanently from paid employment.

Results

Labour Market Transitions and Flow Rates in Australia

Figure 1 indicates the proportion of respondents in each of the nine labour market states in 2001. Almost 60 per cent of adult men were in full-time work compared with just over a quarter of women. Eleven per cent of women were working part-time and caring for dependent children, and a similar proportion caring for children and not working at all. In comparison, only 0.7 per cent of men were in each of these two categories. Women comprised 95 per cent of those caring for dependents and either working part time or not at all.

[FIGURE 1 OMITTED]

The percentage of people who changed their activities between 2001 and 2002 (flow rates) is shown in Figure 2. Of those in full-time work, for example, 88 per cent were still in full-time work a year later (but not necessarily in the same job, or for the whole year).

[FIGURE 2 OMITTED]

Around 69.7 per cent of the sample were in the same labour market situation at both times, and around 30 per cent made a transition to another situation. Around one-third of transitions were outside the labour market, for example' ceasing to care for a child but not entering the workforce. Two-thirds of all transitions (corresponding to 20 per cent of all adults) were either within the labour market (e.g. from unemployment to full-time work) or either into or out of the labour market (e.g. from studying and not working to studying and working part-time). In other words, around one-fifth of the adult population went through some type of labour market transition over a one year period between 2001 and 2002.

The most common transitions were:

* from full-time work to part-time work (other) (6.3 per cent of all transitions)

* from part-time work (other) to full-time work (5.9 per cent)

* from part-time work (caring) to part-time work (other) (4.0 per cent)

* from unemployment to full-time work (3.0 per cent)

* from part-time work (study) to full-time work (2.6 per cent)

* from full-time work to unemployment (2.3 per cent).

The situation with the highest flow rate is unemployment: 70.5 per cent of those who were unemployed in 2001 were no longer unemployed in 2002. People in part-time work were also highly likely to make a transition, with up to 50 per cent of these changing over the one year period. One clear gender difference was that 37 per cent of women underwent a transition compared to only 23 per cent of men.

Men were more likely to move from full-time work to retirement or unemployment whereas women moved into caring roles with part-time or no work. Men were also more likely to move from part-time work to full-time, whereas women were more likely to cease caring and retain part-time work or move out of the labour force completely. The pattern suggests a preference by women for part-time work, regardless of caring status.

Forty-two per cent of the adult population was in full-time work in 2001, and of these, 12 per cent changed their situation during the one year period between 2001 and 2002. Figure 3 shows the transition from full-time work in 2001 to other activities in 2002. The dominant destination for both genders was part-time work (other, with neither caring nor studying orientation). Women were more likely than men to move to part-time work with caring responsibilities (19% compared with 4.5%), and more likely than men to leave the labour force in order to undertake caring duties (11% compared with 1.9%). More men than women moved into unemployment (19.2% compared with 9.1%), and men were also more likely to switch from full-time work to retirement and other non-labour force activities.

[FIGURE 3 OMITTED]

Predictors of Labour Market Exclusion

From the flow rates presented in the first part of this paper, it is clear that there is considerable dynamism in the Australian labour market and that a significant section of the population goes through some type of transition each year. This leads to the question as to whether some transitions are better than others, and questions who is at risk of poor or 'exclusionary' transitions. Exclusion was operationalised in two ways. The first was flows into unemployment. However, since unemployment is basically a measure of point in time exclusion, the second measure of exclusion examined the proportion of the year spent in paid work. This latter figure may be a better measure of long-term exclusion from employment.

This analysis examined the factors which were correlated with labour market participation in the financial year 2001-02. Given these definitions, this section focuses on two key questions:

* Which groups are most likely to move into unemployment?

* What are the key determinants of exclusion from paid employment?

Flows to Unemployment

The first question related to risks of unemploymentand focused on exploring the factors which contributed to greater risk of unemployment in 2002 of those who, in 2001, were in the labour market. Logistic regression was used to estimate the risks of labour market exclusion. The independent variables included in the analysis are shown below. The outcome variable was being unemployed or not at Wave 2. Some categorical variables with more than 2 categories (education, marital status, region) were dummy coded into two separate variables as described:

* Age

* Age squared

* Sex (female versus male)

* Marital status (divorced, widowed or separated; never married; comparison group is married people)

* Level of education (less than Year 12; Year 12 but less than a degree; comparison group is people with tertiary degrees or higher levels of education)

* English proficiency (poor English versus other)

* Aboriginal or Torres Strait Islander status (Indigenous versus other)

* Dependent children (children under 5 versus other)

* Choice in participation (involuntary exit from employment in last 10 years versus other)

* Working in 2001 but wished to reduce their hours (versus others)

* Employment status in 2001 (unemployed versus others)

* Region (major urban area; regional area; comparison group is rural and remote areas).

The results are presented in Table 1. The model predicted 95.2 per cent of cases correctly (Nagelkerke R Square = 0.26). The odds ratios in column two indicate the relative risk of unemployment in Wave 2.

It can be seen that the most significant factors affecting likelihood of unemployment in 2002 were: having been retrenched in the previous 10 years, being unemployed in 2001, being separated or never married, being of Indigenous background, having less than Year 12 education or having a long-term health problem.

Predictors of Exclusion From Employment

The HILDA dataset includes a derived variable which estimates the percentage of time in the previous year spent in paid employment (full time or part time). This variable is specified as continuous variable with values from 0 to 100 per cent of the year. However, 10 per cent of the sample spent no time in paid work and another 75 per cent spent 100 per cent of the year in work. Since these two categories accounted for 85 cent of the total sample, it is apparent that the data violated the assumptions of a normal distribution required for standard multiple regression analysis. The variable was re-categorised into a dichotomous variable with the values of 0 (50 per cent or more of the year spent in employment) and 1 (less than 50 per cent of the year spent in employment).

The sample was people below retirement age in 2002 and who were either in the labour market in 2001 or marginally attached to the labour market, leaving a sample size of 8,424 people. Those not in the labour force but marginally attached were included since they express a desire to work, and are often likely to enter the workforce given favourable circumstances. Inclusion of this group allows us to investigate involuntary exclusion from employment.

Of the sample for this analysis, 84 per cent worked half or more of the 2001-02 year, with the remaining 16 per cent spending less than half the year in work (this category is referred to in the text as 'excluded' since about three-quarters did not work at all during the year).

Logistic regression was used to estimate the risks of labour market exclusion, using the sample independent variables as for the previous analysis but excluding the two variables concerned with employment status in Wave 1.

The results of the analysis are shown in Table 2. The model predicted 85 per cent of cases correctly (Nagelkerke R Square = 0.25). The odds ratios shown in column two indicate the relative risk of working less than half of the year for the predictor in column one (compared to the reference group for each category). In other words, this analysis indicates those who worked from among the population, which could be considered the extended labour force.

Using the population weights included with the HILDA database, there were an estimated 1.7m people in the excluded category between 2001-02. The results show that: people who had been involuntarily unemployed in the last 10 years were five times more likely to be excluded, those with less than year 12 education, with long-term health problems or with poor English proficiency were about three times more likely to be excluded and individuals with children under five years of age, women and Indigenous people were about two-and-a-half times more likely to be excluded. People who were separated or never married had about two thirds higher risk of exclusion. Region had no relationship with levels of employment, but this may be because very broad categories were used. Further analysis showed that, although single status did contribute to less likelihood of employment, the greatest effect was for single people with young children (i.e. sole parents).

Looking only at those in employment in 2001, Figure 4 shows the percentage of each group who spent less than half the following year in employment or, in other words, made an 'exclusionary transition'.

[FIGURE 4 OMITTED]

This shows a similar pattern of risk factors as the regression analyses discussed above with two additional findings: those employed on a casual contract were much more likely to move into exclusion over the following years and those employed in permanent jobs were much less likely to be excluded.

Policy Implications

There are clearly some groups at much greater risk of exclusion from employment, however this is measured. These include those who have been retrenched within the last decade, those with less than year 12 education (and to a lesser extent those with Year 12 but less than university qualifications), people with long-term health problems, migrants with poor English, parents with children under five years of age (especially sole parents), women and Indigenous people.

It is possible that the higher risk for these groups represents the result of screening practices by employers rather than a lack of capacity to undertake employment. This has probably been true over the last decade where there has been a large supply of under-utilised labour. However, some of these factors also represent reduced capacity to undertake employment; for example migrants with poor English would probably be unable to undertake a wide range of jobs.

Since these groups are either already in the labour market or marginally attached, it can be assumed that they would like to work, given the right circumstances. Employer groups are arguing that Australia has an emerging, if not actual, skills shortage especially in some industries (e.g. Dabkowski 2005), and Prime Minister John Howard announced after the Federal election in 2004 that tackling the skills shortage would be a priority for his re-elected government (Howard 2004).

The results of this study have shown that a major (although not sole) determinant of labour market exclusion is lack of initial education, or skills which have become less relevant to changing labour market demands. These results suggest that improved education and training are an important component of policies to prevent or overcome risk for this groups. This direction follows Esping-Andersen's argument that combining flexibility and security requires a 'mobility guarantee' to ensure that some groups are not permanently excluded or marginalised (1999). He emphasises the importance of education and training as part of this guarantee.

The following section suggests three specific policy directions consistent with the TLM framework which would lead to greater participation in employment of those who are currently excluded: a youth guarantee, retraining for workers, and linked employment and training programs for the long-term unemployed.

Policy Directions

A Youth Guarantee

The Dusseldorp Skills Forum (DSF) estimates that over 200,000 young people each year between 15-19 are in the 'at-risk' category of being neither in fulltime work nor full-time study (Dusseldorp Skills Forum 2004). DSF suggests the adoption of youth transitions policies modelled on pilot programs such as that run by the Brotherhood of St Laurence (MacDonald 1999). These policies would aim to provide Year 12 or an equivalent qualification for all young people, and provide more intensive support for early school leavers to either re-engage with education or to gain employment.

A national approach to such a policy could be jointly funded by State and Commonwealth governments. State governments have primary responsibility for secondary education and could commit to providing funding equivalent to 12 years of education for all young people who leave school early. The Commonwealth also has a clear role, if for no other reason than investment in young people at this stage having the potential to save significant future spending on social security and employment programs.

Funds could be pooled into case management programs run by schools, not for profit agencies or others. Such programs currently operate--for example the Jobs Pathways Program funded by the Commonwealth government--but existing resources are inadequate and staff caseloads too high to provide the intensive support necessary.

If additional funding was provided, these programs could use a budget-holding case management model to allow spending on other services to address the young person's needs (for example family counselling, assistance with housing or other training courses and so on). Such a budget would provide more flexibility than at present where case managers often struggle to find services which are accessible or affordable. Programs would have performance targets of a Year 12 or equivalent VET qualification, or possibly entrance into full-time employment (or perhaps enrolment in an apprenticeship or traineeship). The timeframes for these outcomes to be achieved would need some thought, since too short a timeframe might well be unrealistic and skew service provision only to those likely to succeed in the allotted time.

Retraining Policies for Workers

Results from this analysis shows that retrenched workers face long-term risks of exclusion from employment. In the short-term this reflects employer strategies for dealing with fluctuating demand. However, the fact that the impact of retrenchment is still evident up to 10 years later suggests that retrenched workers have experience in jobs which are fewer in number and that they have not had the opportunity to develop new skills. The mantra of 'life long learning' is supposed to ensure that workers are continually re-skilled during their working lives to respond to these changes, but in practice, as argued in Hancock's article in this volume, there is little policy in Australia directed to this end. A major barrier to participating in education for many workers is the inability to find substitute income to allow time off to study--the proposal outlined below is one way in which this could be accomplished.

Individual learning accounts for all workers could be established with the aim of building enough funds to pay for extended periods of study (possibly one or two years). Funds would be contributed by employees, employers and government, with the largest government contribution going to those at most risk of retrenchment, or the lowest paid. As an example, around $60,000 would be required for a low-paid worker to take two years off to study. Over twenty years (and not including interest accrued from savings), this would require savings of $3,000 per annum or $60 per week. This would be too great to expect individual workers to pay.

Governments could subsidise half the cost for low-paid workers, with subsidies tapering off as income increased. Employers and employees could share the other half, perhaps as part of a one-off deal around wage adjustment as occurred with superannuation contributions. Employees could contribute their notional wage increase to the learning account, on the condition that employers made a matching contribution. Alternatively, existing entitlements such as long-service leave, could be folded into the accounts. It might be possible to identify occupations and industries at greatest risk of skills obsolescence through risk analysis of time series data, or alternatively basing contributions solely on income.

Funds would be held by a public or private industry body so that entitlements were carried over from one employer to the other. The CoINVEST building industry fund is an example of an institution which operates in this manner for long-service leave entitlements. Alternatively, funds could be invested with private superannuation companies, but a limit would have to be set on fees and charges to prevent interest being swallowed up for small accounts.

The learning account would mainly fund income replacement. Fees for courses could be paid through a HECS-type mechanism, but some care would be needed to ensure that workers did not accrue large debts which might act as disincentives for participating. If TAFE fees of around $1200 per year were charged, it would probably only take one or two years to pay off a HECS debt, but university course fees would of course take much longer.

Some type of approval process would be necessary to ensure that accounts were used only for accredited qualifications, and not misused as has occurred in the UK (Haukka, Keating and Lamb 2004). A minimal process might be a simple check on whether the proposed course was accredited, while a more extensive process might include a requirement that workers have a 'training plan' developed by a career counsellor. Finally, learning accounts could be rolled over into superannuation funds if not used before retirement age.

This type of individualised savings approach might be criticised as overly complex. Why not simply establish a government transfer payment, and fund it through increased personal and company taxation? This would have some advantages, both in terms of simplicity, and possibly in equity considerations for at-risk groups of workers. A disadvantage of the learning account model is that those with least ability to save will face the highest retrenchment risk. A savings model might also discriminate against women who take time off work to care for children.

One of the attractions of an individual approach is that it builds legitimacy as individuals can see they are reaping some benefits themselves. Moreover, a savings scheme might encourage workers to think about the use of their accounts to study, more so than they may have considered otherwise. The high degree of public support for hypothecated taxes such the Medicare levy and for the tripartite-funded superannuation system suggests that a contributory approach could attain much greater public support than a social security program, and may therefore be more politically viable. However it would still require a significant public contribution to redistribute funds from low-risk to high-risk workers, occupations and industries.

Linked Employment and Training Programs for the Long-Term Unemployed

The current employment services provided by the Job Network take a 'work-first approach which mainly emphasises job search and related activities. For the majority of jobseekers, this approach seems reasonably effective, but for long-term unemployed people, it has been of less assistance. The Brotherhood of St Laurence (BSL) has developed an innovative employment program to assist long-term unemployed public housing tenants into employment (Temby, Housakos and Ziguras 2004). The target group includes adults receiving unemployment payments, people with long-term health problems and sole parents. Most have not worked for at least two years before entering the program.

Case management and intensive personal support is used to assist people to manage personal issues (such as health problems or family conflict) whilst at the same time undertaking work-oriented activities. Supportive pre-vocational training builds a combination of work and personal development skills such as communication, team building and conflict resolution. These skills are crucial to the successful transition from unemployment to sustainable employment. The training enables people to establish daily work routines, manage their personal issues while maintaining their training/work placement schedule and confirm their interest in employment.

Work experience provides people with the opportunity to 'learn on the job' within a supportive work environment, increases self-esteem and confidence, re-acquaints long-term unemployed people with a daily work routine and the expectations of an employer and provides the experience required when applying for work. The BSL provides work experience through pre-vocational training and by linking people into traineeships in such areas as office administration, maintenance, childcare and aged care. Trainees gain full certificates and crucial support is continued on the job.

Post-placement support is provided to people in the period following the start of a new job. Post-placement support for both the employee and the employer greatly enhances a smooth transition to work and continuing employment. It is also provides a 'safety-net' for people who become unemployed again and offers further support with access to alternative employment, training or support service options. An important aspect of this program is that trainees are employed by a Group Training Company (GTC) run by the BSL, rather than by employers. The GTC deals with employment costs and performance issues and takes some of the risk away from employers in hiring a long-term unemployed person.

Overall, the BSL program appears to be relatively successful in reintegrating those at greatest risk, and has some similarities to the principles proposed for the Neighbourhood Renewal program in the UK (Campbell 2001).

Conclusions

The results of this study show that the Australian labour market is very dynamic, with around 20 per cent of the adult population each year making a transition into, out of or within employment combined with other activities such as study or caring. Part-time work plays an important role in many transitions. The transitional labour market framework is therefore a useful analytical tool for studying Australian working conditions. Kruppe (2002) reporting on the period 1994-95 for European Union countries, found around 16 per cent of the population went through a transition each year, although this figure is not directly comparable due to the different categories of transitions used. However, he reports that around one-third of transitions took place outside the labour market, a similar figure to that found here.

In the HILDA Australian data, women were more likely to undergo a transition during the year than men, again reflecting results from European countries (Kruppe 2002). Women are far more likely to be undertaking care for a young child (around 95 per cent of carers either working part-time or not working are women) and to move into and out of caring roles combined with part-time work. Women appear to have a general preference for part-time work, whereas men appear to prefer full-time work. For example, of those working part-time and caring for a young child, 45 per cent of women who change their situation stay in part-time work but cease caring for a child, but around the same proportion of men move to full-time jobs.

Research on European welfare state regimes (Muffels, Wilthagen and van den Heuvel 2002) found that Liberal regimes (UK and Ireland in that study) combined a high level of labour mobility and flexibility with a low level of work security. Although the figures are again not directly comparable due to different time periods of the studies, our results certainly appear to support the contention that Australia reflects this pattern. Further work to develop measures of transitions consistent with European studies would be useful to allow greater comparative analysis.

There are clearly some groups at much greater risk of exclusion from employment. These include those who have been retrenched within the last decade, those with less than year 12 education, people with long-term health problems, migrants with poor English, parents with children under five years of age (especially sole parents), women and Indigenous people. These findings reflect similar results from previous studies (such as. Le and Miller 2001, Wilkins 2004) and overseas evidence (Campbell 2001).

This paper finds evidence of a segmented labour market with a peripheral group with low skills and in oversupply. On the other hand, employer groups have over the last year argued that they face difficulties recruiting workers with adequate skills and experience for their industries. Any strategy to overcome exclusion from employment must therefore include a significant training component, and for most groups this will mean vocational education. Policies which combine training and work experience are likely to be effective for long-term unemployed people and young people at risk of early school leaving. A new policy focus on income replacement for workers to retrain while still in the workforce, would be a more effective 'early intervention' approach compared with the current situation, and may help to address looming skill shortages.

This paper has shown that the TLM framework is a useful one for describing the current labour market in Australia. However, systematic risks of exclusion for a significant number of people under the current policy settings, suggest a need for new directions in managing flexibility and risk. The TLM approach appears to have great potential for conceptualising new models to ensure that citizens are able to make positive and productive transitions.

References

Australian Bureau of Statistics (2004), Australian labour market statistics April 2004, Cat. No. 6105.0, Australian Bureau of Statistics, Canberra.

Buchanan, J. and Watson, I. (2000), 'Beyond the wage earner model', in Reshaping Australia social policy: changes in work, welfare and families, Growth 48, (ads) L. Hancock, B. Howe, and A O'Donnell, CEDA, Melbourne.

Burgess, J., Mitchell, W. and Preston, A. (2003), 'The Australian labour market in 2002', The Journal of Industrial Relations, vol. 45, no. 2, pp. 125-50.

Campbell, M. (2001), New deal for communities: national evaluation scoping phase (worklessness evidence review), Department for Transport, Local Government and the Regions. Available at: http://www.renewal.net/Nav.asp?Category=:worklessness [accessed 19 October 2004].

Cass, B. (1988), Income support for the unemployed in Australia: towards a more active system, Australian Government Publishing Services, Canberra.

Castles, F. (1996), 'Needs-based strategies of social protection in Australia and New Zealand', in Welfare states in transition, (ed.) G. Esping-Andersen, Sage, London.

Cebrian, I., Lallement, M. and O'Reilly, J. (2000), 'Introduction', in Working-time changes: social integration through transitional labour markets, (eds.) J. O'Reilly, I. Cebrian and M. Lallement, Edward Elgar, Cheltenham, UK.

Chou, Y.K. (2003), 'The Australian growth experience, 1960-2000: human capital, R and D, or steady-state growth?', Australian Economic Review, vol. 36, no. 4, pp. 397-414.

Dabkowski, S. (2005), Skills shortage only brake for industry, The Age Online, 5 January 2005. Available at: http://www.theage.com.au/new/Business/Skills-shortage-only-brake-for- industry/2005/01/04/1104832108572.htm?oneclick=true [accessed 2 March 2005].

Dusseldorp Skills Forum (2004), How young people are faring: Key Indictors 2004, Dusseldorp Skills Forum, Glebe, NSW.

Esping-Andersen, G. (1999), Social foundations of postindustrial economies, Oxford University Press, Oxford.

Giddens, A. (1998), The Third Way: renewal of social democracy, Polity Press, Cambridge.

Haukka, S., Keating, J. and Lamb, S. (2004), Alternative mechanisms to encourage individual contributions to vocational education and training, National Council for Vocational Education Research, Adelaide.

Howard, J. (2004), 'The fourth Howard Ministry', media release 22 October 2004. Available at: http://www.pm.gov.au/news/media_releases/media_Release1134.htm [accessed 24 February 2005].

Keating, M. (2003), 'The labour market and inequality', Australian Economic Review, vol. 36, no. 4, pp. 374-96.

Kruppe, T. (2002), 'The dynamics of employment in the European Union: an exploratory analysis', in The new dynamics of full employment: social integration through transitional labour markets, (eds) G. Schmid and B. Gazier, Edward Elgar, Cheltenham, UK.

Le, A.T. and Miller, P.W. (2001), 'Is a risk index approach to unemployment possible?', Economic Record, vol. 77, no. 236, pp. 51-70.

Macdonald, F. and Holm, S. (2001), 'Employment for 25- to 34-year olds in the flexible labour market: a generation excluded?', in Future directions in Australian social policy: new ways of preventing risk, Growth 49, (eds) L. Hancock, B. Howe, M. Frere, and A. O'Donnell, CEDA, Melbourne.

MacDonald, H. (1999), Bridging the gap: assisting early school leavers to make the transition to work, Brotherhood of St Laurence, Fitzroy.Muffels, R.J.A.,

Schmid, G. (1998), Transitional labour markets: a new European employment strategy, Discussion paper FS I 98-206, Social Science Research Centre Berlin. Available at: http://www.wz-berlin.de/default.en.asp [accessed 5 September 2003].

Schmid, G. (2000), 'Transitional labour markets: a new European employment strategy', in Innovative employment initiatives, (eds) B. Marin, D. Leulders and D.J. Snower, Ashgate, Aldershot.

Schmid, G. (2002), 'Towards a theory of transitional labour markets', in The Dynamics of full employment: social integration through transitional labour markets, (eds) G. Schmid and B. Gazier, Edward Elgar, Cheltenham, UK.

Standing, G. (2002), Beyond the new paternalism: basic security as equity, Verso, London.

Temby P., Housakos, G. and Ziguras, S. (2004), Helping local people get jobs: Insights from the Brotherhood of St Laurence experience in Fitzroy and Collingwood Brotherhood of St Laurence, Fitzroy. Available at:

Watson, N. and Wooden, M. (2002), The Household, Income and Labour Dynamics in Australia (HILDA) Survey: Wave 1 survey methodology, Melbourne Institute of Applied Economic and Social Research, Melbourne.

Wilkins, R. (2004), 'The effects of disability on labour force status in Australia', Australian Economic Review, vol. 37, no. 4, pp. 359-82.

Wilthagen, T. and van den Heuvel, N. (2002), Labour market transitions and employment regimes: evidence on the flexibility-security nexus in transitional labour markets, Wissenschaftszentrum Berlin, Berlin.

Ziguras S., Considine, M., Hancock, L. and Howe, B. (2004), From risk to opportunity: labour markets in transition, Centre for Public Policy, University of Melbourne, Melbourne. Available at: http://www.public-policy.unimelb.edu.au/research/ TLMs_background_paper.pdf.

Stephen Ziguras, Centre for Public Policy, University of Melbourne

Table 1: Predictors of Unemployment in 2002

                            Odds
Predictor                   Ratio    p

Involuntarily lost
  job in last 10 years      6.45    ***
Unemployed in 2001          5.74    ***
Divorced, separated
  or widowed                2.17    ***
Indigenous                  2.17    ***
Less than Year
  12 education              1.82    **
Never married               1.67    **
Long term health problems   1.54    **
Migrant with poor
  English proficiency       1.38    ns
Year 12 education but
  less than degree          1.20    ns
Female                      1.16    ns
Age squared                 1.00    ns
Age                         0.97    ns
Has children under 5        0.97    ns
Major urban region          1.10    ns
Other urban region          0.98    ns
Was working in Wave
  1 but wanted less work    0.95    ns

Source: Household, Income and Labour
Dynamics of Australia (HILDA) survey,
Waves 1, 2001 and 2, 2002.

Table 2: Predictors of Exclusion from Employment

                            Odds
Predictor                   Ratio    p

Involuntarily lost job
  in last 10 years          5.24    ***
Less than Year
  12 education              3.39
Long term health problems   2.89    ***
Migrant with poor
  English proficiency       2.84    ***
Has children under 5        2.66    ***
Female                      2.51    ***
Indigenous                  2.40    ***
Divorced, separated
  or widowed                1.76    ***
Never married               1.64    ***
Year 12 education but
  less than degree          1.58    ***
Age squared                 1.00    ***
Age                         0.83    ***
Major urban region          0.96    ns
Regional area               0.90    ns

Source: Household, Income and Labour Dynamics
of Australia (HILDA) survey, Waves 1, 2001
and 2, 2002.