3  What are low wages?

Berstein and Hartman (1999:28) provide a very useful summary of what constitutes the low wage labour market. They observe that there are two approaches to characterising this market.

Job-based definitions identify a set of jobs characterised by low wages, few benefits, and little upward mobility. Worker-based definitions are typically based on a worker’s absolute or relative hourly wage, earnings (wages times hours worked), or educational level. Job-based definitions provide the theoretical foundation and worker-based definitions, the empirical basis for study of the low-wage labor market.

The job-based approach is strongly associated with the concept of segmented labour markets. Segmented labour market theory argues that there is not a single market where people are smoothly allocated to whatever job best matches their abilities, and paid accordingly. Rather, there are two distinct markets. The primary market has well paid jobs, with reasonable levels of security and opportunities for skill development and advancement. The secondary market, in contrast, comprises low paid jobs that are often insecure, have few fringe benefits or promotion possibilities and little opportunity for workers to improve their skills and wages. A key element of the concept of segmented labour markets is that it is difficult for workers to move from the secondary market to the primary market: there is much more job mobility within each of the segments than there is between them.[1]

The extent to which the concept of a segmented labour market is illuminating remains controversial in economics. It is difficult to find good empirical data on the features that are distinctive to this view of the labour market. In this review, we focus on the worker-based definition of the low wage labour market. We discuss later the extent to which low wage workers are indeed stuck in low wage jobs. But we do not consider, as does segmented labour market theory, the roles of discrimination and other systemic features in trapping people into low wage jobs.

The definition of what constitutes a low wage is arbitrary. The choice made is important because it substantially affects both the numbers and characteristics of low wage workers.

There are three broad approaches to defining a low wage. The first enquires into the minimum earnings needed to ensure that an individual or family of defined composition is able to live at an acceptable standard of living. This was the approach adopted by Justice HB Higgins when he determined the first basic, or minimum, wage in Australia in 1907. The difficulty with this approach is in determining an acceptable standard of living for a full-time worker and his/her family. Of course, a given wage will deliver a different material living standard to families of different size and composition, so a wage that is adequate for a single young person may not provide an acceptable income for a family of five. These are some of the compelling difficulties that have caused the Australian Industrial Relations Commission to abandon the attempt to identify a “living wage” in recent decisions that set minimum wages.

A second approach to defining a low wage is to relate it to administratively-set expressions of what constitutes a low but acceptable income. This might be a legislated minimum wage, or some multiple of a selected social welfare payment, such as an old age pension. Such an approach enables the researcher to avoid imposing his or her own interpretation of what is a low income, and to substitute instead that determined by the parliament or some agent of government. The drawbacks with this approach are twofold. If the minimum wage is used, it will often be the case that only a small number of people (especially adults) are actually receiving this rate. If some multiple of a social welfare benefit is used, then the choice of multiple (to reflect the disutility of work) becomes as arbitrary as selecting a minimum acceptable standard of living.

The third approach defines a low wage as some fraction of the median wage. Again, the choice of fraction is arbitrary. Two-thirds of the median full-time wage is a frequently used figure. This makes low wages a relative concept – there is no direct relation with the ability to purchase some minimum basket of goods and services. A related, even more relative, approach is to take the wage that defines the bottom quintile or decile of wage earners. If a given percentile is used to define low wages then it follows that the proportion of the workforce that is employed on a low wage cannot change.

When choosing among these options, it is useful to select a measure that is widely used, in order to facilitate comparisons across countries and studies. Two-thirds of the median wage of a full-time worker is probably the most widely used measure, but the empirical work to be reported below is not uniform in the definition adopted. The two-thirds median was adopted by the European Commission Working Group on Equitable Wages. It has the advantage of being somewhat above the minimum wage of most countries, which means that a non-trivial number of workers will be encompassed within the definition. It should be noted that the conclusions about who is a low paid worker, and the degree of mobility in and out of this state, are somewhat sensitive to the choice of low wage. Specifically, levels of mobility are higher, the lower is the value of the wage chosen. So is the proportion who are youth.

People can have low earnings because they are not employed full-time - working for only a fraction of the normal working week or only a fraction of the normal number of weeks in the year. There is also quite a deal of short term volatility in the level of wages earned, especially among new entrants to the labour force. It is preferable to include as low wage workers only those who have persistently low weekly earnings, arising from a low hourly wage. People with low wages have low earnings, but people with low earnings do not necessarily have low wages. They may be receiving quite reasonable wages, but voluntarily or involuntarily, be working less than a normal working week or year. We therefore mostly exclude from consideration the issue of low earnings, as distinct from low wages.

The links between low wages, low earnings and low standard of living are set out in Figure 3, below.

Low wages will lead to low earnings. But people who have adequate wages may also have low earnings, if they work a sufficiently small number of hours per week or per year. People with low earnings may in turn have a low standard of living. But it is quite possible, and indeed common, that they do not. The link between the level of earnings and the standard of living is indirect, because people mostly live in families where resources are shared. If the low wage/low earnings worker lives with one or more others who have adequate earnings, then his or her standard of living is likely to be quite adequate. We show a dotted link between adequate earnings and low standard of living. People who have adequate (or not-low) earnings may have a low standard of living, if they or their family have particularly high expenditure needs (such as a large number of dependent children, or family members with a disability).

The relationships illustrated in Figure 3 may hold at the current time (eg, for a week, month or year) or they may hold over a more extended period. The major policy concern with low wages is that people can have a low standard of living for many years, caused by low earnings, that are in turn caused by low wages. This is a sub-set of the people who at any moment of time have low wages, and is not the whole cause of a low standard of living.

There is some empirical work that examines the link between the relationships set out in Figure 3.

Figure 3 - The relation between low wages, low earnings and low standard of living

Savage (1999) quotes a study by Dixon for New Zealand that finds that about one quarter of low wage earners live in households that are in the bottom third of the household income distribution (p 8). Richardson (1998) and Richardson and Harding (1999) conclude that, for Australia, while the majority of low wage workers do not live in low income households, a sizeable minority does. Again, the experience is not homogeneous. “There are indeed people who earn low wages who also live in comfortable middle and upper income households. There are also low-wage workers who support dependent children and who struggle on low incomes.” (Richardson, 1998:576-7). For Ireland, Nolan (1998) finds that while 24% of full-time employees are low-paid (earn less than two-thirds of the Irish median gross weekly earnings), as few as 6 and 13% of them live in households below half and 60% of average equivalent income, respectively. These proportions of equivalent incomes are commonly used as poverty lines. However, the overwhelming majority of full-time employees in poor households (as many as 90% in the UK) are low paid. So while the experience of poverty implies being low paid for people who are full-time workers, the reverse is often not true. Similar outcomes are found in the United States (eg, see Burkhauser and Finegan, 1993).

In exploring the link between low wages and low standard of living, the OECD (1997:12) observes that there is a high positive correlation, across countries, between the incidence of low paid employment and the proportion of people living in low income households. They report that while in most European countries, 10% or fewer of full-time low wage workers live in poor households (ie, with an income of less than half the average adjusted household income), in the US the comparable figure is about 25%.