2.2  Developments in the wage structure and low wage work

English-speaking countries have had similar developments in their wage structures over the past 25 years. While the trends have been the same, the degree of change has been different. The changes have generally been most marked in the United States, followed by the UK. They have been more muted in Canada, and particularly in Australia. The European OECD countries have had a more varied experience. For ease of exposition (and because they have been the most extensively examined), we here summarise the main developments in the US.

• Wage inequality has risen substantially since the early 1970s for both men and women full-time employees, with particularly fast growth in wages occurring at the very top.
• Real wages for men fell over the two decade period to the mid-1990s, with the lowest wage earners losing the most (up to 30% of the value of the real wage).
• Increased cross-section wage inequality has not been offset by a rise in wage mobility over time or a compression in non-wage benefits, so that longer term inequalities have also risen.
• Wage differentials by education, occupation and age have all increased, but the gender differential has decreased. At the same time, wage inequality within age, education, sex and occupation groups has risen.
• Since the mid-1990s, real wages have increased their rate of growth and all points on the wage distribution have benefited.

The rise in inequality in the wage structure in the US has been dramatic and has translated into a substantial rise in the inequality of household incomes and in consumption. For lower skill men, there has also been a fall in employment and in the value of the real wage. This has occurred while the overall levels of education have risen considerably. Bernstein and Hartmann (1999), for example, report that men who had not completed secondary school had real hourly wages that were on average 30% lower in 1997 than in 1973. The comparable figure for women was 3% lower. Both men and women who had completed secondary school but not done post school education had falls in their real hourly wage of 16%.Over the 32 year period ending in 1995, the real weekly wage for men in the bottom 30% of the wage distribution fell up to 5%. The higher the wage, the faster the wage growth. The ratio of the wages of fully employed men at the 90^th percentile to that of men at the 10^th percentile of the wage distribution rose from 3.3 in 1963 to 4.7 in 1995. (Katz and Autor, 1999:1468,1471,1475).

Table 1 shows the considerable difference in the course of male wage inequality across the OECD. The big increases for the US and UK stand out (followed by the increase for New Zealand). Austria, France, Norway, Sweden, Finland, Germany and the Netherlands by contrast have had only small rises in inequality, or even a fall in the case of Norway and Germany.

These patterns are suggestive of an important role of differences and changes in labor market institutions and regulations in explaining the cross-country divergence of wage structure changes in 1980s and 1990s – And the existence of either a decline in the relative wages of the less skilled, a sharp rise in the unemployment of the less skilled, or both in almost all OECD countries over the past two decades despite expanding relative supplies of highly educated workers is strongly suggestive of a common shift in labor demand against the less skilled.

Katz and Autor, 1999:1503-4.

There is widespread agreement among labour economists that one aspect of “labor market institutions and regulations” that has an impact on wage inequality is the level and enforcement of a minimum wage. There is clear evidence for the US, UK, Australia, New Zealand and several European countries that changes in the level of minimum wages are directly inversely correlated with the level of wage inequality. Where the real value of minimum wages has been allowed to fall, overall wage inequality has risen: the more minimum wages have fallen, the more inequality has risen. (Blau and Kahn, 1999:1434; Keese, Puymoyen and Swain, 1998:235).

Note: Ratio of wage of 90^th percentile earner to 10^th percentile earner

Source: Katz and Autor, 1999:1503

As Figure 1 shows, the upper earnings limit (expressed in PPP $US) of full-time workers in the 10^th and 20^th percentiles of the wage distribution do not vary a great deal across a range of OECD countries. Low wage workers are paid particularly well in Switzerland, Germany and the Netherlands, but there are only small differences among the remaining countries. New Zealand low wage workers have the lowest wages of those reported.

There is much greater variety in the extent of low wage employment among the OECD countries than in the level of wages of low paid workers. Figure 2 shows the proportion of full-time workers in a range of OECD countries who were, in the mid-1990s, receiving less than two-thirds of median earnings. The proportion varies from a high of 25% for the US to a low of 5% for Sweden. It is notable that it is English-speaking countries (and Japan) that have a high incidence of low pay. This includes New Zealand, with 16% of its full time workers earning less than two thirds of median earnings.

Figure 1 - Low earnings of full-time workers 1994: annual gross earnings, expressed in US$ using purchasing power parities for private consumption

Source: Keese, Puymoyen and Swaim, 1998:225

Figure 2 - Incidence of low pay in selected OECD countries: 1994/5

Source: OECD 1997:48. Low earnings are defined as less than two thirds of median full-time earnings. Data are for full-time workers.

2.3  Implications for upward wage mobility

There is a growing literature on the phenomenon of a so-called “skills-biased” technological change. This bias in technological growth in favour of high skills has widened the wage gap between the low-skilled/low-paid and the highly-skilled/well-paid employees. That is, low wage jobs pay less than they used to in real terms and relative to the average. This real decline in the low skills wage could be due to changes in the characteristics of the low skills jobs and workers, or to a fall in the returns to low-skills jobs. If the latter, it suggests that the proportion of non-training jobs has risen. Gregg and Wadsworth (2000) argue that there has been a widening gap between the pay levels of entry level jobs and other jobs in Britain. Further, this is attributable to the characteristics of, and not the rewards to, low-skills jobs and low skilled workers. They conclude that there has been “---a simple decline of [the pay of] entry jobs in the distribution of all jobs unrelated to observable characteristics and returns to those characteristics.” (p 516).

It is important to appreciate that, in the English-speaking world and especially in the US, the rise in the gap between the wages of low skilled workers and high skilled (educated) workers has arisen in large part because of the fall in the wages of the less skilled/educated. One reason for this is, as Gregg and Wadsworth identify, that the average ability of young people with low levels of education is most likely falling. As the average levels of education of a cohort rise, those who are left behind (drop out of school) are not a random sample of the group. They are those who for a variety of reasons do not find formal education beneficial.

While technological change might have been increasing the productivity of high skill workers more than that of low skill workers, demand for each has grown at a comparable pace. That is, technological and other changes have not seen low wage jobs disappear: to the contrary, they have grown in proportion to the high paying jobs, at the expense of middle-paying jobs. This phenomenon has been referred to as the “disappearing middle”.

Freeman (1998), among others, gives evidence of what he called a “bifurcated growth” of employment in the US. This uneven growth is evident in both occupations and industries, with both high-wage services and low-wage services growing faster than any other industry group in the US. The US Bureau of Labor Statistics forecasts that of the ten jobs expected to provide the largest jobs growth to 2006, seven of them require only “short-term on-the-job training” as the required skill. They include retail salespersons, truck drivers, home health, teaching and nursing aides, and receptionists and information clerks. The other three jobs with expected fast employment growth are at the high skills end, requiring at least a college degree. The seven low skilled jobs (and there are many others like them also forecast to expand their share of employment) do not require much skill to perform and do not provide promotion ladders. It is most unlikely that workers who stay in these jobs would benefit from much skills development or wage mobility. Furthermore, this phenomenon seems to be generalised among OECD countries. If this trend continues, the fears of a low-wage/low-skills trap for individuals may prove both well founded and worsening over time, as it would be harder to cross the ever-widening hollow middle to make it to the high-wage end.