RBA Annual Conference – 1998 Introduction Guy Debelle

While in recent years the unemployment rate has fallen from its peak of over 11 per cent in the early 1990s, the current rate of unemployment – at just over 8 per cent, about the average for the past fifteen years – is still of concern for both economic and social reasons. From an economic perspective, unemployment represents the underutilisation of one of the economy's main resources, labour. Socially, unemployment is associated with an array of problems, not least a lower standard of living and lower self-esteem for the unemployed.

The higher rate of unemployment since the mid 1970s is not a problem unique to Australia. A large number of OECD countries have experienced a similar rise. There have, however, been some exceptions. Most notably, the unemployment rate in the United States has fluctuated around a relatively constant level for over three decades. Whereas in the 1960s and early 1970s the structure of the labour market in the United States was sometimes criticised for delivering higher unemployment rates than other OECD countries, today it is often held up as the example to which other countries should aspire. On the other hand, the relatively low unemployment rate in the United States has been associated, in the past two decades, with widening wage inequality.

The contrast between continental Europe and the United States is often seen as illustrating a trade-off between a lower unemployment rate and a more unequal earnings distribution. An important consideration in such a trade-off is the interaction of the labour market with the social security system. A wider distribution of labour income could be mitigated by an appropriately targeted tax and welfare system that sought to avoid high effective marginal rates of taxation that might act to discourage individuals from seeking or accepting employment opportunities.

The papers in this volume were commissioned by the Bank and the Centre for Economic Policy Research at the Australian National University to consider the issue of unemployment in Australia and to contribute to the debate about policies to reduce unemployment. The papers covered three main issues:

• The labour market experiences of other OECD countries have been diverse over the past three decades, as evidenced by the range of unemployment outcomes. These varied experiences can provide lessons for Australia about the impact on unemployment of different labour market institutions and the effectiveness of different policies. Cross-country comparisons can also shed light on the existence of the ‘diabolical trade-off’ between higher unemployment and greater wage inequality, and on measures that can be undertaken to circumvent such a trade-off.
• The microeconomic aspects of the Australian labour market are important in determining the appropriate policies to reduce unemployment. However, little is known at the microeconomic level about the way in which employment and wages adjust in response to developments both within and outside the labour market. Nevertheless, it is clear that the incidence of unemployment has been unevenly distributed across different sections of the Australian population.
• There is an array of possible solutions to unemployment. A common theme that runs through all the papers presented at the conference is that there is no easy policy prescription to reduce unemployment. Rather, any approach must encompass a range of different measures, and further, important synergies may emerge when a coherent strategy is adopted that simultaneously implements these measures.

International perspectives

Many commentators have highlighted the distinction between labour market developments in continental OECD Europe and the United States. The stylised description is that the United States has enjoyed a lower unemployment rate than continental Europe because of greater flexibility in its labour market, but that this has come at the cost of a wider dispersion of incomes.

The paper by Jackman reveals that such a stylised description conceals a variety of differences within continental Europe. The dispersion of unemployment rates between European countries is greater than that between Europe and the United States. Some of the countries with relatively heavily regulated labour markets, such as Austria, have had low unemployment rates for a number of years.

Furthermore, a simple static comparison between Europe and the United States may be misleading. An important feature of labour market institutions is their ability to cope with unexpected shocks. In the 1960s, the different structures of labour market institutions across countries may have been less important in the relatively benign economic climate. However, the more tumultuous economic environment of the 1970s may have exposed the previously hidden deficiencies in labour market structures in a number of countries.

It is difficult to isolate those features of the US and the various European labour markets that have led to such divergent outcomes. Moreover, it is dangerous to draw strong conclusions about the efficacy of a particular labour market policy without considering the interactions of that policy with the other labour market institutions in the country concerned. Thus, the adoption in Australia of a policy that has been effective in another country may not generate the same beneficial outcomes given the different institutional structure.

This lesson is further borne out in the comparison between the labour market experiences of Australia, New Zealand and the United Kingdom described in the paper by Wooden and Sloan. Significant differences in the speed and process of labour market reforms in the three countries have not, to date, generated dramatically different labour market outcomes. However, the labour market outcomes are also likely to have been affected by other factors, such as differences in the macroeconomic environment and the nature of the reform process in other areas of the economy, in particular the product market.

The United States is often held up as the benchmark for assessing the effectiveness of labour market institutions in other developed countries. The paper by Katz describes the multiple dimensions of the US labour market experience. From the standpoint of overall employment outcomes, the US labour market is outstanding. The employment to population ratio in the United States has risen steadily over the past twenty-five years, indicative of the fact that the US labour market has absorbed the concurrent large increase in labour supply, particularly of females. The US labour market has provided good outcomes for females and young workers compared with a number of European countries where labour market interventions have served to primarily protect the employment of prime-aged males at the expense of other groups in the workforce. On the other hand, the employment to population ratio of prime-aged males has declined in the US.

The degree of, and trends in, income inequality in the United States appear less favourable in international comparisons. The US has historically had a relatively wide wage distribution and over the past two decades, wage inequality has been growing and real wages at the bottom end of the distribution have fallen. This outcome has been partially attributed to the impact of skill-biased technological change; that is, developments in technology over a number of years have tended to favour high-skilled or more-educated workers.

It is often argued that the interaction of skill-biased technological change with the more flexible wage-setting system in the US has resulted in increased wage dispersion, whereas in many European countries, skill-biased technological change has been reflected in rising unemployment rates of less-skilled workers. However, the empirical evidence suggests that such a characterisation is simplistic. Firstly, unemployment rates are higher in Europe than in the US across all skill categories. Secondly, much of the rise in inequality is ‘within group’ inequality; that is, wages have become more dispersed for workers with very similar characteristics including education and measurable skills. The lack of a satisfactory explanation for these trends in inequality in the US complicates the task of drawing lessons for Australia.

Unemployment in Australia

Since 1960, the ratio of the number of people employed to the working-age population in Australia has fluctuated around a relatively constant level, while the unemployment rate has risen from around 2 per cent in the 1960s to an average of around 8½ per cent over the past fifteen years, with most of the rise occurring in the second half of the 1970s. By definition, this implies that the growth in the demand for labour has not kept pace with the growth in the supply of labour, which is indicative of shortcomings in the operation of the Australian labour market.

The trends in the unemployment rate over the past thirty years highlight the key influence of economic growth and trends in labour costs on labour market outcomes. Increases in the unemployment rate occurred primarily in three relatively short episodes associated with the sharp contractions in economic activity in the mid 1970s, 1982–83 and 1990–91. Furthermore, the rises in the unemployment rate in these episodes tended to be considerably faster than the falls in the subsequent recoveries, highlighting the costs of variability in growth. The sharp rise in labour costs in 1974 and 1982 exacerbated the cyclical rises in unemployment experienced at those times (and indeed contributed to the contractions in economic activity themselves). Falling unemployment throughout the second half of the 1980s owed a lot to restraint in aggregate wage growth, as well as the sustained period of expansion in economic activity.

The effects of adverse movements in labour costs may also be quite persistent. While restraint in the growth of wages has seen real unit labour costs return to the levels of the 1960s, the unemployment rate still remains considerably higher. This is consistent with the idea that changes in labour costs can have long-lasting effects through their influence on investment decisions which, in turn, determine the stock of physical capital and therefore the productive potential of the economy.

Examining developments in the labour market from an aggregate perspective can conceal important developments at a more disaggregated level. The paper by Borland and Kennedy documents variations in the incidence of unemployment across different groups in the population. While the unemployment rate has risen for all sections of the population, the rise has not been uniform. The labour market experiences of males and females have differed significantly: the male employment to population ratio has declined by more than 15 percentage points over the past thirty years, whereas that of females has increased by a similar amount. This largely reflects the divergent performance of industries which predominantly employ males, such as manufacturing, compared with those which employ more females, such as service industries. It also reflects the large increase in part-time employment. Over the medium term, one might expect to see the divergent labour market trends of the two genders diminish as males increasingly seek part-time work and employment in industries that have historically employed more females.

The other notable divergence in labour market outcomes is that between skilled and unskilled workers. Unemployment rates are considerably higher for less-educated and less-skilled workers. Again, this is not particular to Australia, but is evident in most OECD countries, and may reflect the impact of skill-biased technological change. As discussed above, skill-biased technological change need not lead to higher unemployment rates for less-skilled or less-educated workers if the relative wage paid to those workers declines. In Australia, the tendency has been for quantity adjustment rather than price adjustment, as evidenced by the relatively higher unemployment rates of the less skilled. However, there has also been a quality adjustment in the Australian labour force. The increase in high-school retention rates and the increased participation in tertiary education both suggest that people are responding to the divergence in the unemployment rates induced by the skill-biased technological change, by seeking to improve their skill and education levels. If this is so, then in the medium term, the increased supply of skilled labour should reduce the difference in unemployment rates and decrease wage dispersion.

The substantial rise in the unemployment rate and the wide range of unemployment rates across different groups in the population both suggest that much of the adjustment in the labour market occurs, at least in the shorter term, through adjustment in quantities – that is, the number of people employed – rather than price – that is, the wage. The institutional structure of the labour market has a large influence on the extent to which prices or quantities adjust.

The institutional structure of the labour market and the social security system also have a large influence on income distribution. The unemployed receive relatively low incomes because their primary source of income is unemployment benefits. Thus, the unemployed are at the lower end of the individual income distribution. In principle, however, many unemployed people could live in households where there are other sources of income which might offset their low personal income. Harding and Richardson present evidence in their paper that suggests this is generally not the case. Many unemployed people live in a household where there is no other breadwinner. Consequently, households with an unemployed member are disproportionately concentrated at the lower end of the household income distribution. In contrast, households with individuals who are employed on relatively low wages are more evenly spread across the household income distribution.

The evidence also shows that government cash benefits are the primary source of income for families with an unemployed member, but are only a very small source of income for families with wage and salary earners, implying that the welfare system in Australia is relatively well targeted. However, it also means that any possible solution to unemployment must consider the interaction of changes in the wage structure with the welfare system.

Finally, to understand the rise in unemployment it is important to understand how the labour market adjusts to adverse developments. It is necessary to know how firms adjust their hiring and investment decisions in response to changes in the macroeconomic environment and to changes in the costs of employing labour or capital. The paper by Freebairn documents that, unfortunately, there is comparatively little evidence on the microeconomic workings of the Australian labour market. Such information is vital when considering the design of policies to address unemployment.

Solutions

When contemplating possible solutions to unemployment, the experiences of Australia and other OECD countries indicate no obvious or easily implementable path. Rather, an effective solution needs to draw on a number of different elements. Furthermore, consideration of the interaction of the different components of the policy package is important, rather than analysing and adopting each policy in isolation. For example, labour market programs are likely to be more effective in an environment of sustained growth.

A critical element in any approach is the maintenance of a non-inflationary rate of economic growth that is as steady as possible. The paper by Dungey and Pitchford highlights that output growth is a major determinant of employment growth, and that recessions can have large and long-lasting detrimental effects on the unemployment rate. Consequently, macroeconomic policy can contribute to improved labour market outcomes by aiming to achieve the highest possible rate of economic growth while maintaining a low inflation environment.

There are, however, limits to the extent to which growth alone can permanently reduce the unemployment rate. Unemployment in Australia currently has a sizeable structural component. Attempts to further reduce unemployment below this rate through macroeconomic stimulus are likely to be inflationary. Specific policies are needed to reduce this structural component.

The paper by Debelle and Vickery suggests that moderate, but sustained, wage restraint can deliver sizeable reductions in the structural unemployment rate. The difficulty is determining the means by which to achieve aggregate wage restraint. In abstract terms, too high a level of aggregate wages reflects an imbalance of labour market power between active participants in the labour market (insiders), and those with a less effective presence (outsiders). Such an imbalance can diminish the benefits of productivity-enhancing reforms if the increase in productivity is captured predominantly by those already employed, in the form of higher wages, rather than being more evenly distributed across the whole population, in the form of increased employment.

One means of increasing the bargaining power of outsiders is to use active labour market programs. Drawing on the large variety of programs that have been adopted in OECD countries, the paper by Martin presents a checklist of those labour market programs that appear to work and those that do not. The general aim of such programs is to increase the ability of the unemployed to compete effectively in the labour market by increasing ‘job readiness’ in terms of basic work skills and by assisting the unemployed in locating employment opportunities through the provision of information, or the provision of wage subsidies which increase their attractiveness to potential employers. Labour market programs which target potentially disadvantaged groups at an early age, while initially expensive, can avoid greater costs in the future.

A particularly effective type of labour market program is one that provides opportunities for the workforce to increase their education and skill levels. However, care must be taken to avoid the pitfall of encouraging individuals to remain in certain forms of education which, while temporarily reducing the measured unemployment rate, do not provide them with the necessary training to increase their future employability. Education, like all other forms of labour market programs, needs to be carefully targeted. Furthermore, labour market institutions should be structured to allow the unemployed with increased skills and education to compete effectively with the insiders.

A necessary element in any solution to unemployment is entrepreneurship. Without a sufficient pool of entrepreneurial skills, there will not be adequate employment opportunities for the workforce to utilise their skills and education. Impediments to risk-taking need to be minimised so that businesses can create jobs in response to enhanced labour market flexibility and the availability of skills.

Some decrease in unemployment could be achieved through increased relative wage flexibility, rather than through increased aggregate wage flexibility. That is, the wide range of unemployment rates across the different groups in the workforce may reflect the inability of the wage-setting system to allow relative wages to adjust to imbalances in labour demand and supply. Allowing greater adjustment in relative wages may also deliver a lower aggregate wage level, as relative wages fall for those groups experiencing the higher unemployment rates.

An outcome where wages fall for certain sections of the population may raise concerns about equity. However, the inequity of lower wages must be weighed against the inequity of unemployment, particularly given that unemployed households are, in general, lower in the income distribution than low-wage households. An important issue in such a debate is whether the employment generated by the lower wage outcomes is the first point in a career path that leads to higher wages later in the working life, or whether those gaining employment predominantly remain locked into low-wage employment. The evidence on this in Australia is, however, fairly scant.

A policy prescription to reduce unemployment that includes a reduction in real wages should also consider the interaction of the wage system with the social welfare and tax systems. The social welfare system is an effective channel for providing assistance to low-income households, but it is essential to ensure that it does not create incentives for benefit recipients to avoid participation in the workforce. Thus, a necessary component of any reform package should be the reduction of high effective marginal tax rates facing lower income workers, without compromising the social objectives of the welfare system.

In summary, there are many dimensions to unemployment and correspondingly many elements in any solution to unemployment. A policy approach that is likely to lead to a sustained fall in unemployment will require the maintenance of steady non-inflationary growth, and will include measures that increase the employability of those without jobs, such as education and re-skilling, and increase their relative bargaining power, to restrain growth in labour costs.