Introduction

To be in paid work but poor used to be a contradiction in terms in Australia's 'wage earners' welfare state'. In the early 1970s, the Commission of Enquiry into Poverty estimated that less than two per cent of families with an adult in full-time employment could be described as poor (Burbidge, 1981). Rather, poverty was mainly a problem for those who could not get waged work.

In the 1990s, however, having employment is no longer a guarantee of staying out of poverty. Changes in the structure of employment and growing inequalities in the distribution of wages seem to be nudging Australia in the direction of a US-style population of 'working poor'. If this is the case, the political fallout could be substantial, since a significant element of the support attracted by One Nation appears to stem from the anxieties of people who feel themselves stuck on the margins of poverty in spite of their paid labour.

The ACTU's Living Wage Case attempted to put working poverty on the public agenda by expressly linking the question of low pay with that of household deprivation (Buchanan and Watson, 1997). Yet many low-paid workers live in better-off households, and the links between individual and family income status are complex.

Certainly there is greater inequality in the distribution of earnings. The Accord in the 1980s led to a fall in real wages at the lower end and a widening of the gap between low and high incomes. The subsequent shift towards enterprise bargaining and individual agreements have seen this process continue. Since 1993, people in the top quartile of pay outcomes under enterprise agreements have received annual increases of at least five per cent, whereas those in the bottom quartile have received only two to three per cent on average (Department of Workplace Relations, 1998). The lowest paid, of course, tend to rely on minimum Federal or State awards. During the 1970s, these moved in line with the 'benchmark' Henderson poverty line for a family of four, but in the 1980s they began to fall below it (ACOSS, 1997).

Family income, however, is not determined only by the primary earnings of individuals. It is also affected by family composition and the number of earners. Women's participation in paid work has grown substantially in recent decades, although much of the increase has come in part-time work, while that of men has dropped. Yet the opportunities for women have not expanded evenly across socioeconomic groups and there is growing polarisation between women in and out of the labour force (Probert, 1997). This is also contributing to a polarisation between dual-earner families and those with only a part-time earner (or no earner at all).

Taxes and social security benefits also have a major impact on family income. Part of the trade-off negotiated under the Accord in return for wage moderation was an increase in the social wage. Benefits targeted on families with children were boosted in the late 1980s, because of a recognition that child poverty had been on the increase and family payments have continued to expand both in scope and value in the early 1990s (Eardley, 1997). The net effect has been largely to maintain lower-paid working families' disposable incomes in real terms, but without moderating the gains in market earnings achieved by higher-paid families (Landt and Beer, 1998). Thus inequalities in net family incomes have also widened.

A study in progress at the Social Policy Research Centre suggests that a trend towards working poverty has accelerated in the 1990s, and this article draws on the research to trace the links between individual low pay and family poverty. The first question is what is happening to wages. Are more people getting low pay and if so who are they?

Trends in Low Pay

There is no straightforward definition of 'low pay'. Wages are both payments for work carried out and a means of subsistence. Pay can therefore be seen as 'low' relative to the work involved or relative to the worker's needs. For comparison over time and with other countries, the measure used here is a relative one, based on an hourly rate calculated as total gross weekly pay divided by the number of hours normally worked each week. The threshold for low pay is taken as two-thirds of the median hourly rate for all waged workers (see Eardley, 1997, 1998, for further discussion of low pay measurement issues).

As Australia currently has a system of 'junior' wages payable to people under 21, a lower threshold has been applied to them. Here we use a junior low pay threshold of 52.6 per cent of the adult threshold, based on the ratio established in the AIRC 1997 wage case ruling, which set a minimum hourly wage of $5.00 for juniors ($9.50* 0.526).

Table 1 shows the trend in receipt of low pay. Contrary, perhaps, to expectation, there does not seem to have been any substantial overall change in the rate for adults, although the absolute numbers have increased in line with the growing work force. The trends are quite different for men and women, however. Low pay for men has been rising steadily, while for women it has been declining, such that their risk of being in low pay has more than halved compared to that of men. The absolute number of women with low hourly wages grew because of their increased labour market participation, but men still made up three-quarters of the total increase (from around 460 000 persons in low pay in 1981-82 to 660 000 in 1995-96). In spite of the relative improvement in women's hourly rates of pay, however, the proportion with low levels of weekly earnings remains considerably greater than that of men because women are more likely to work part-time.

Since by definition half of all employees must have rates no higher than the median, Table 1 implies that just under 40 per cent of all adult employees have hourly pay rates clustered between two-thirds of the median and the median itself. This means that in spite of a widening earnings gap overall, Australia continues to have a relatively flat earnings distribution at the lower end. Even if young people are not assigned a lower threshold, the proportion of employees with low hourly pay only rises to 14 per cent in the mid-1990s, compared with 22 per cent in the UK in 1994, based on the same relative measure (Webb, Kemp and Millar, 1996).

What is more striking is that the proportion of young people in low pay even on the basis of a junior wage threshold appears to have more than doubled in the 1990s, having been declining up to then. There are some uncertainties about the comparability of SIHC data on young people with that from earlier surveys since new data collection methods were introduced in 1994-95, but one explanation may lie in the shrinkage of the pool of young people in the labour market and not in education or training.

The proportion of young people aged 15-19, not in education but in the labour force has dropped from just over half in 1979 to only 28 per cent in 1998 (ABS, 6203.0, various dates). By contrast, the proportion in education and also in the labour force increased from eight per cent in 1979 to 28 per cent in 1998. In the early 1990s there appears to have been a sharp increase in part-time work - often with very short hours and disproportionately concentrated among the lowest paid. This suggests a shift in the availability of youth employment low-paid, part-time, casual work, and is clearly also influenced by greater numbers of students accessing part-time work. Not only are an increasing proportion of young low-paid workers primarily students, but most of them also live with their immediate family or other relatives - nearly 84 per cent of all waged workers aged under 21 and virtually all the young low paid.

Aside from sex and age, the low paid are in many ways similar to other employees. Low-paid workers are less likely to be or to have been married than other workers, which is not surprising given that many are young. Single employees and working sole parents are also more likely to be low paid than members of couples, as are those with few or no qualifications. Interestingly, the proportion of the low paid who were in full-time work rose between 1981-82 and 1995-95, from just under 60 per cent to nearly two-thirds.

Notes:(a)Individuals included are those aged 15-64 (men) and 15-59 (women), receiving current income from employed work.

(b) Hourly pay rates are defined by mid-points in bands of reported working hours

(c) In 1981-82 the SIHC grouped ages 20-24, so for this year 'Junior' includes those only those aged 15-19 and 'Adult' includes those aged 20 and upwards.

Source: ABS Surveys of Income and Housing Costs, unit record files

Part of the change over time in the percentage of employees receiving low pay reflects an industry shift towards lower wage sectors. For example, although the rate of low pay in the retail sector barely changed between 1981-82 and 1995-96, the sector itself more than doubled in size and contributed nearly half of all the extra low-paid workers over this period. The rate of low pay in the 'personal and other services' sector actually fell, from 45 to 19 per cent, but the absolute number of low-paid service workers increased by around 25 000 because the sector expanded from employing less than one per cent of all workers in 1981-82 to nearly four per cent in 1995-96.

One important question is how far low pay is a special problem for migrants and people of non-English speaking backgrounds. This is a matter of some controversy, since there are claims that an ethnic 'underclass' is forming in areas of Australia's cities (Birrell and Seol, 1998). The SIHC itself provides only limited information on ethnicity, but it does show that while still a small group, employees born overseas but outside Europe have nearly doubled as a proportion of the working population over this period. Their hourly pay situation, however, differs little from that of employees generally. In 1981-82 they represented 5.8 per cent of all employees and 5.5 per cent of the low paid, whereas by 1995-96 they made up 9.2 per cent of employees but 10.2 of the low paid.

Low Pay and Poverty

To see how far low pay and poverty overlap, we need to locate low-paid workers in the families where they live. Table 2 gives an indication of the change over time in the relationship between low pay and poverty. It shows the percentage of employees living in ABS-defined income units with equivalent disposable annual incomes below the poverty line, according to whether they were currently receiving low hourly pay rates.

Again, one of the most striking features of this analysis is the apparent surge in the percentage of young people living in poor income units, irrespective of whether they were low paid themselves, especially between 1989-90 and 1995-96. Part of the explanation is likely to be that those who are in the labour market and working are counted as income units in their own right even if they are still living with their parents. As we have seen, the young low paid represent an increasing proportion of a shrinking population and many of them are students. Despite being counted as separate income units many could not afford live on their own incomes alone without some assistance from their families.

The data are perhaps more reliable as an indicator of what is happening amongst adults. In 1981-82, the vast majority (nearly 97 per cent) of all employees did not live in poverty. More than one in ten low-paid workers did, but even amongst them the overlap between individual low pay and family poverty seems to have been small. By 1995-96, however, the rate of income poverty amongst waged workers apparently more than doubled, accounting for around one in five low-paid adult employees. This is a worrying development, but it is still a long way from saying low pay equals poverty. Clearly for a large majority of employees it does not. Much the biggest numerical increase in family poverty took place among those employees who were not low paid (from around 87 000 in 1981-82 to 327 000 in 1995- 96).

Notes:(a)Low pay is as defined in previous table, using separate adult and junior thresholds

(b) Income is annual equivalent disposable income and poverty is determined using the relevant after-housing cost Henderson Poverty Line.

Source: ABS Survey of Income and Housing Costs unit record files

Table 3 further illustrates the complex relationship between individual low pay and family income distribution. It places individual low-paid workers within the quintile distribution of equivalent income units. Thus, in 1989-90, just under eight per cent of all employees and 20 per cent of those with low hourly rates of pay lived in families whose equivalent disposable income placed them in the bottom quintile. Over the first half of the 1990s there was a substantial overall shift downwards into the bottom two income unit quintiles, with the percentage of all employees in the bottom quintile doubling to over 17 per cent. Again, however, we see that this downward shift took place as much amongst those not low paid by our definition as amongst the low paid. While by the mid-1990s a considerably larger proportion of low-paid workers lived in families in the bottom fifth of the income distribution than in 1989-90, nearly 40 per cent were still in the third quintile or above. It seems that individual low pay, while significant, is only one of the factors contributing to the increase in working poverty.

If we look at the characteristics of low-paid individuals in poverty we find that there has been a shift towards men, full-time workers and single people, and away from women, sole parents and couples with children. This is consistent with the earlier finding that men appear to have fared worse in terms of low pay than women.

Notes:(a)Because the quintile groups are based on the individual and income unit sizes vary across equivalent income groups, the numbers in each quintile are not the same.

(b) Low pay for adults and youth is based on their different respective thresholds

(c) Numbers are weighted and rounded to the nearest thousand

Source: ABS SIHC 1989-90, 1994-95 and 1995-96, unit record files

The growth in family poverty amongst full-time workers would seem to belie the proposition that poverty is mainly a problem for part-timers unable to get as much work as they need, unless even full-time work is becoming more casual and intermittent. This is a possibility, since a large proportion of all employment is now only available on casual or contract terms (Burgess and Campbell, 1998). However, the SIHC data show that in 1995-96 the vast majority (93 per cent) of all employees were in work for at least three-quarters of the previous period (however long that was). Just under a quarter of individual employees in families below the Henderson poverty line had less than this level of consistent paid work, compared with five per cent of those not in poverty. This suggests that while there is a relationship between intermittent employment and family poverty, it is not a conclusive one. Saunders (1994) estimated that in 1989-90, families with a head in full-time, full-year work made up around 5.3 per cent of all those with incomes below the Henderson poverty line, up from four per cent in 1981-82. There are some difficulties in replicating the same full-year work indicator in the new SIHC survey, but our estimates suggest that this figure grew to over 10 per cent by the mid-1990s.

Conclusion

The findings reported here rely on judgements about appropriate measures for both low pay and poverty which are, of course, open to question. The analysis has also been restricted mainly to one data source and to a narrow concept of income poverty.

Nevertheless, it seems that the phenomenon of working poverty in Australia is real and is growing. Low pay on an hourly basis does not in itself equal poverty. Most low-paid workers do not live in families with incomes below the poverty line and the biggest increase in family poverty has been among employees not in low pay. Yet the proportion of low-paid workers who are also in poor families is now about one in five. Partly this is due to the increasing prevalence of involuntary part-time and casual work. On the other hand the proportion of employees in full-time work but still poor has also increased.

Letting wages fall at the bottom end of the distribution, as has been urged by business and some economists as a way of boosting employment, seems likely to exacerbate the working poverty problem further, even if the impact would not only be on those in or close to poverty, unless compensatory support is provided through the tax and social security system.

To increase family payments cannot be the whole answer, however. For a start, half the low paid in poverty do not have children. Extending income-related support to low earners without children, whether through social security or through a tax credit, would not only be costly but in the context of increasingly individualised wage bargaining might be captured over time by employers through even lower wages, unless a strong minimum wage platform is in place. Means-tested payments and credits can enhance incentives for unemployed people to take lower-paid work, but they also create poverty traps and can act as a disincentive for women in couples to look for work of their own. A danger is that if wages are allowed to fall on the assumption that family incomes will be protected by social security or tax, the payments may end up failing to meet their income support goals even while spending on them increases to the point where it loses public support.

Looking at it from the other perspective, what happens if we raise minimum wages? Clearly this would affect more than just the poor, while some workers whose individual pay is slightly above the minimum, but whose income unit is still poor, might not benefit. On the other hand, the international evidence suggests that the negative impact of minimum wage increases on employment is exaggerated, for adults at least (OECD, 1998). Combating working poverty in Australia probably requires a combination of strategies, which might include judicious increases in safety net wages and awards affecting workers in low-wage industries, as well as carefully designed support through the tax and social security systems. Perhaps most important would be a recognition that, while not yet at US levels, working poverty in Australia requires serious policy attention if we do not want to build up intractable problems for the future.

References

Australian Industrial Relations Commission (1997), Safety Net Report - Wages, AIRC, Sydney.

Australian Council of Social Service (1997), Living Wage, Paper No. 86, ACOSS, Sydney.

Australian Bureau of Statistics (1997), Average Weekly Earnings, Catalogue No. 6302.0, ABS, Canberra.

Australian Bureau of Statistics (various dates), The Labour Force Australia, Catalogue No. 6203.0, ABS, Canberra.

Birrell, B. and B.-S. Seol (1998), 'Sydney's ethnic underclass', People and Place, 6, 3, 16-29.

Buchanan, J. and I. Watson (1997), 'The Living Wage and the working poor', in M. Bittman (ed) Poverty in Australia: Dimensions and Policies, Reports and Proceedings No. 135, Social Policy Research Centre, University of New South Wales, Sydney.

Burbidge, A. (1981), 'Working people in poverty', in R. Henderson (ed) The Welfare Stakes: Strategies for Australian Social Policy, Institute of Applied Economic and Social Research, Melbourne, 147-174.

Burgess, J. and I. Campbell (1998), 'Casual employment in Australia: growth , characteristics, a bridge or a trap?', Economics and Labour Relations Review, 9, 1, 31-54.

Commission of Inquiry into Poverty (1975), First Main Report: Poverty in Australia, AGPS, Canberra.

Department of Workplace Relations (1998), Wage Trends in Enterprise Bargaining, March Quarter.

Eardley, T. (1997), 'Does Australia Have a Problem of Working Poverty?', in P. Saunders and T. Eardley (eds) States, Markets, Communities: Remapping the Boundaries, Proceedings of the National Social Policy Conference, Volume 1, Social Policy Research Centre Reports and Proceedings No. 136, SPRC, University of New South Wales, Sydney, 103-124.

Eardley, T. (1998), 'Low pay and Poverty', Social Policy Research Centre Newsletter No. 68, February, SPRC, University of New South Wales, Sydney.

Landt, J. and G. Beer (1998), The Changing Burden of Income Taxation on Working families in Australia, Discussion Paper No. 29, NATSEM, University of Canberra, Canberra.

Organisation of Economic Cooperation and Development (OECD) (1998), 'Making the most of the minimum: statutory minimum wages, employment and poverty', Employment Outlook, July, OECD, Paris.

Probert, B. (1997), 'The social shaping of work: struggles over new boundaries', in P. Saunders and T. Eardley (eds) States, Markets, Communities: Remapping the Boundaries, Proceedings of the National Social Policy Conference, Volume 1, Social Policy Research Centre Reports and Proceedings No. 136, SPRC, University of New South Wales, Sydney, 5-27.

Saunders, P. (1994), Welfare and Inequality: National and International Perspectives on the Australian Welfare State, Cambridge University Press, Melbourne.

Webb, S, M. Kemp and J. Millar (1996), The Changing Face of Low Pay in Britain, Bath Social Policy Papers No. 25, University of Bath, Bath.

Australian Institute of Family Studies, Level 20, 485 La Trobe Street, Melbourne, Vic 3000, Australia. Telephone: (03) 9214 7888; International: 61 3 9214 7888. Facsimile: (03) 9214 7839; International: 61 3 9214 7839. URL: http://www.aifs.gov.au/