Skills shortages: concepts, measurement and policy responses (1).

Abstract

Understanding the different concepts and causes of skills shortages is important if sound policies are to be developed in the areas of employment, education, training and skill formation. This paper outlines different perspectives of skills shortages held by economists, employers and unions. It then discusses the measurement and identification issues and provides an example illustrating the practical difficulties of assessing skill shortages. A number of policy options to deal with skills shortages are also discussed.

Introduction

The concept of a skills shortage has different meanings to different people. Public discussions often have as a sub-text the view that governments should 'do something' about skills shortages. However, this is not always appropriate and, even when it is, it is necessary to be able to define and measure the shortage prior to responding to it.

Employers, employees and policy analysts have varying perspectives on what skills shortages mean. The lack of a common understanding of the concept often obscures analyses of problems, their causes and possible solutions. According to Arrow and Capron (1959), some proposals for solving shortage problems stem from a misunderstanding of the causes of shortages as well as from an exaggeration of the evidence.

A level of anxiety currently exists about skills shortages in Australia, and elsewhere, and its impact on the economy. Shortages could result in under-utilisation of capacity and consequently reduced production. If the response of employers is to bid up wages in ways that do not match productivity then higher inflation could result. Another effect of bidding up wages in times of shortages is that the existing supply of workers circulates more quickly between employers, thus making retention more difficult and the employees involved less productive. From workers' perspective, a shortage, or even a perception of it, can be a good thing because it delivers higher wages. It is against this background that this paper reviews the issues relating to skills shortages.

Section 2 defines skills and discusses the different perspectives on and causes of skills shortages. A review of the various ways skills shortages can be identified and measured is presented in section 3. Section 4 briefly outlines the program of monitoring of skills shortages as currently undertaken by the Department of Employment and Workplace Relations (DEWR) and reviews recent studies of ICT skills shortages in Australia. Section 5 contains some policy responses to skills shortages and the last section contains some concluding remarks.

Perspectives on and causes of skills shortages

This section looks at the different meanings that are often attached to the term skills shortage in public discussion and the causes of skills shortages. In many such discussions the issue of the types of skills individuals should possess for different types of work often arises and for this reason we begin by a brief overview of what are skills. The discussion below draws on a number of previous studies (Bosworth, Dutton and Lewis 1992; Roy, Henson and Lavoie 1996; Borghans, de Grip and van Smoorenburg 1998; Veneri 1999; Frogner 2002; National Skills Task Force 2000) which have provided the theoretical framework for analysing skills shortages and the associated practical problems.

What are skills?

A skill is an ability to perform a productive task at a certain level of competence. As a skill is associated with a particular task, a person who does not possess such a skill is unlikely to be able to carry out this task or will be less productive than somebody who does possess it. Skills are often associated with a qualifications and their acquisition through formal education and training. An individual can, however, acquire skills in other ways, including various forms of informal learning and on-the-job experience. A hierarchical structure for skills is often used with each skill's placing determined by the level of knowledge needed and the autonomy of decision-making involved in completing the set of tasks. As a job involves doing a number of tasks, to perform in a job a person must possess a bundle of skills.

Becker (1962) distinguished between general and specific skills. (2) He defined general skills as those that are useful to many firms. Skills that are useful to only a few firms, particularly to those firms providing the training to acquire them, are called specific. Completely general training increases the marginal productivity of trainees by the same amount in all firms while completely specific training has no effect on the productivity of trainees in all other firms except the one providing the training. Where a worker acquires general skills, an employer may need to increase his pay or risk losing him to another employer. For this reason, according to Becker, firms will not pay for the training for general skills because they consider employees to be the main beneficiaries from such training. (3)

The training and education literature is primarily concerned with general skills, that is, skills useful to more than one employer. This literature often distinguishes between basic, generic and vocational/occupational skills.

* Basic skills include literacy, numeracy and, increasingly, computer literacy as the foundation skills necessary for the development of advanced skills (National Skills Task Force 2000). The public usually pays for either the total cost or else a very significant part of the cost of providing these skills and their benefits extend beyond the labour market to personal development and social capital.

* Generic skills are a range of skills including problem solving, team working and the ability to improve personal learning and performance.

* Vocational skills are technical skills needed to carry out tasks within an occupation or an occupational group. They are transferable across firms. Vocational training provided on-the-job (eg apprenticeship training) is usually paid for in part by the employee, in terms of lower wages than could be obtained elsewhere and quite frequently by the government as an employer subsidy.

A fourth type of general skill is what the National Skills Task Force (2000) calls 'personal attributes', which some argue (eg Keep and Mahew 1999) are not really skills. They are often defined in terms of motivation, judgement, leadership and initiative. Employers often cite them as attributes that they look for when hiring new workers. The analysis of Employer Manpower and Skills Survey in Britain in Green, Machin and Wilkinson (1998) showed that some employers valued motivational and attitudinal 'skills' very highly and indicated a deficiency of these skills among their employees.

Perspectives on skills shortages

Here we first discuss some of the views of the main players on skills shortages and then define three possible meanings applied in practice to this term.

Economist perspective

From an economist's or 'market' perspective a skills shortage occurs when there is insufficient supply of appropriately qualified workers willing to work under existing market conditions, particularly the prevailing wages. In other words, the reference is to the external labour market, the aggregation of all firms' demand and all individuals' supply of the specific skill. Sometimes the shortage is evident only in particular specialisations of an occupation or in a particular geographical location. Unfilled or hard-to-fill vacancies are indicators of a shortage.

Over time, the market adjusts in a number of ways, including price and/or quantity adjustment or substitution with other types of labour, and the imbalance clears.

Employer perspective

Employers' perspective on skills shortages is important because often they are the main, if not the only, source of data allowing judgements to be made about the existence of a skills shortage in the market. The employer perspective is, however, usually in terms of recruitment difficulties experienced by individual employers. These do not necessarily lead to unfilled vacancies in the short run but can result in other labour market difficulties or rising labour costs that the employer may regard as a shortage.

The analysis of an employer-based survey of skills and skills shortages by Green, Machin and Wilkinson (1998) shows that although many employers equate skills shortages with hard-to-fill vacancies, a substantial number do not. The analysis showed that some employers viewed internal skills deficiencies, or gaps, in the same light as skills shortages. Others were declaring the existence of skills shortages when in fact they were unwilling to offer competitive wages--a situation that is discussed below.

In the short run, instead of recruiting new workers, some firms react to an increase in the product demand with a range of coping mechanisms that could involve the reallocation of resources within the internal labour market. This is because of the time lags involved in recruitment from the external market. A firm might increase hours of work per employee by reducing short time or increasing overtime or offer to convert part-time contracts to full-time, which incidentally increases capital utilisation. It might change the incentive system to increase worker effort and thus improve efficiency (Akerlof & Yellen 1986; Solow 1979). When a firm attempts to adjust employment to meet new product demand, then its response shifts to interaction with the external labour market. The hiring costs increase disproportionately with the time required to make appointments. In a tight labour market, if the costs of making an appointment within a certain time are high some employers view this as a skills shortage.

If the expansion in product demand is unsustainable, then additional costs in terms of firing costs could be involved, although in a highly deregulated labour market such costs could be minimal. Therefore, in such a situation, some firms will initially only hire new labour on a casual basis, and only when there is more certainty in the demand for the product sustaining will they consider putting new employees on ongoing contracts.

Irrespective of whether the firm relies on the internal or external labour market to solve the problem of increased product demand, if costs increase at the margin then some firms may view this as a skills shortage. In a competitive market where the firm has to sell its product at the going price, it may not be profitable for it to meet the whole of the potential increase in the demand for its product. Only expansion in output by other existing firms can meet the increased product demand. Each firm might view this as a shortage because the higher output would be profitable if they had sufficient employees working at 'normal' hours, wage levels and effort.

With continuing problems, some firms substitute skilled labour in short supply with differently qualified labour, which could include the recruitment or redeployment of less skilled workers and work reorganisation. Sometimes this necessitates additional training, either internally provided or purchased from an external provider. In such situations, even though the employer might feel a skills shortage exists, from a market perspective the positions were filled and hence there is no shortage.

To sum up: an employer may regard as a shortage some recruitment or labour cost issues that have not resulted in an increase in unfilled vacancies.

Union perspective

Skill is central to the bargaining between labour and management over wages (Rainbird 1992). Thus, it is expected that the views of unions on skills shortages, in terms of quantity and quality of labour, would be at variance with those of employers. Unions generally want to improve the skill level of their members because higher skills increase their bargaining power in wage negotiations. The union perspectives on skills shortages, however, are not uniform and depend on the union's strength and the nature of the labour market in which it operates.

In occupations with strong union representation, such as policing, nursing and teaching where demand is sometimes set against some desired staffing standards, it would appear to be in the interest of the union to overstate the extent of current and future skills shortages because it helps push up the wages of new entrants and increase union membership. It is hoped that there would be a wages flow-on for the benefit of already employed members. However, this may not be the only union objective. Often the union's perspective of shortages is that of a desired level of staffing seen as necessary to offer better quality services.

The organisation and monopoly power of some unions and professional associations enable them to restrict supply and assist in creating scarcity of skills. Through certification and lengthy periods of training, entry into the union and the occupation is controlled. For example, some medical specialist groups have significant powers over who will be admitted for training and the numbers that will be trained (Borland 2002; Seldon, Jung and Cavazos 1998; Friedman and Kuznets 1945).

Differentiating meanings of skills shortages

It is necessary to have appropriate terminology to distinguish the various meanings attached to the term 'skills shortage'. Based on DEWR (2004b p.12), three meanings are distinguished:

Hard-to-fall vacancies

Hard-to-fill vacancies exist when employers are unable to fill or have considerable difficulty filling vacancies for an occupation (or specialised skill needs) at current levels of remuneration and conditions of employment, and reasonable location. Hard-to-fill vacancies are those that are still unfilled after a reasonable period. DEWR and others use the term 'skills shortages' to describe this but we refer to it as hard-to-fill vacancies, which is what is actually observed, and to distinguish it from other meanings.

Hard-to-fill vacancies are typically for specialised and experienced workers, and can coexist with relatively high overall unemployment in the occupation. Hard-to-fill vacancies are aggregated across firms to gauge the extent of the problem across the market.

Skills gap

Skill gaps occur where existing employees lack the required qualifications, experience and/or specialised skills to meet the firms' skill needs for an occupation. Skill gaps may apply to new employees, where employers are unable to find suitable applicants for an occupation and recruit workers who need further training and/or experience to meet the firm's skill needs for the occupation.

Skills gaps do not necessarily relate to formal vocational qualifications relevant to an occupation, as sometimes the complaint is about generic skills (eg computer literacy) or attitudes (eg customer relations).

In some instances skills gap may be latent. This can occur when employers do not perceive that there is a problem because they are not fully aware of skills needed for optimal production.

Another aspect of a skills gap is one that Roy, Henson and Lavoie (1996) call the normative skills gap. It is the gap between what is and what should be. For example, a government's target of a particular nurse/patient ratio in public hospitals is a benchmark against which to assess skills shortages.

Recruitment difficulties (4)

Recruitment difficulties occur when employers have some difficulty filling vacancies for an occupation. There may be an adequate supply of skilled workers, but employers are still unable to attract and recruit sufficient suitable employees.

These difficulties may be due to characteristics of the industry, occupation or employer, such as: relatively low remuneration, poor working conditions or image of the industry, unsatisfactory working hours, location hard to commute to, ineffective recruitment or firm-specific and highly specialised skills needs.

Recruitment problems are not widespread enough for them to be declared a market-wide shortage. With appropriate offers of wage compensation workers could be enticed to work in many of these situations described above.

Causes of skills shortages

There are three main causes of skills shortages following a shift in demand. These are slowness in the adjustment of wages, slowness in the adjustment of supply and inadequate timely labour market information.

Slowness in the adjustment of wages

Much of the theory about defining and identifying occupational shortages developed from research that focussed on wage movements (Blank and Stigler 1957; Arrow and Capron 1959). (5)

The price-quantity graph in Figure 1 provides a useful way to explain the role of wage adjustment in the dynamics of skills imbalance. It shows the initial demand curve for a particular skill at D1 and the supply curve at S. The market is at equilibrium at point el with wages We and supply of skills at Qe. Suppose that the demand for the skill increases from D1 to D2. In this case, unless the wages adjust instantaneously to the new equilibrium value of Ws, the quantity of skills supplied will be less than the quantity demanded by the amount that is the difference between Qmax and Qe. Hence, under current market conditions, employers will be unable to hire workers with these skills and a shortage will occur.

[FIGURE 1 OMITTED]

In practice, however, wage adjustments of this kind usually take time. Impediments include delays by employers in accepting the need for wage adjustment and the reluctance to disturb existing wage structures in order to raise the rates for new employees with the required skills (Arrow & Capron, 1959). Some of the factors that tend to reduce the market clearing function of wage adjustments and increase its stickiness are institutional and regulatory arrangements, such as contracts of employment between employers and employees, imperfect product market competition, lack of transparency of the market, wage controls and social welfare provisions (OECD 1994).

Arrow and Capron (1959) introduced the concept of a dynamic shortage in which demand for a skill grows faster than supply. They explained the cause of dynamic shortage as rapid and persistent rise in demand, low elasticity of supply and a slow market 'reaction speed', which is defined as the ratio of the rate of wage rise to the excess of demand over supply. Slow reaction time stems from a number of factors. Firstly, it takes time for firms to recognise the existence of a shortage at the current wage rate. If the initial recruitment effort fails, they often try the strategy of increasing advertising expenditure in the hope of attracting suitable applicants. This strategy is likely to fail if the equilibrium wage has indeed increased. Secondly, it takes time for the firm to decide on the level for a new wage rate and the number of employees to hire at that rate.

Another impediment to wages adjusting could be the existence of a wages agreement within the firm, industry or the whole economy such as the 'Accords' in Australia in the 1980s and early 1990s. Employers' reluctance to offer higher wages to attract new recruits stems from a fear that higher salaries for new recruits will have a flow on effect on salaries of existing staff. In such situations, the firm may provide in-house training either to lesser skilled employees in the firm or to new hires without the required skills to prevent wage inflation.

In some sectors, the government influence on the supply of and demand for skills is significant, and the management of the market is more direct. Education and health are two sectors where the government has significant influence on the management of the market. In such cases, achievement of market equilibrium will be difficult to achieve because wage adjustments might respond very slowly to changes in demand or supply. In contrast to a dynamic model, this is an example of a static model.

Skills shortages can occur when the equilibrium wage for a particular skill is below some minimum level. Here the minimum wage could correspond to the wage equivalent of public welfare benefits payments (Roy, Henson and Lavoie 1996). Figure 2 depicts one of these situations, typically encountered in low-skilled occupations. Initially, the equilibrium wage is We with demand D1, but no worker is willing to work at this wage because it is below the 'socially acceptable minimum wage' of Wm. Under this situation, employers will report skills shortages or recruitment difficulties. If employers are allowed to hire temporary or permanent workers from outside the country's labour market, eg guest workers, the supply curve shifts to $2, and equilibrium is achieved at el but only if the socially acceptable minimum wage is inapplicable to the guest workers.

[FIGURE 2 OMITTED]

Slowness in the adjustment of supply

The speed at which supply adjusts depends on a number of factors. The main factors are the length of training required and the responses of the training institutions to changes in demand. Quite often education and training requirements take a long time to complete, although refresher courses for those who are returning to the same occupation after an extended period of absence from it can accelerate the process. Supply might vary because of changes in education and training provision and changes in preferences for various forms of work and to demographic changes including ageing, emigration and immigration.

Educational institutions adjust only slowly in offering places for particular courses. The time taken for training authorities to recognise a particular need, the internal budgetary processes in public vocational education and training and higher education institutions and the current staffing patterns mean that the number of places provided in areas that are in shortage adjust slowly over time.

The expansion of publicly funded places is not the only factor affecting the numbers. Students must be attracted to take up the new places, and in the case of apprenticeships secure contracts with employers. A decrease in student demand for places can occur when individuals traditionally trained in a given area have other more attractive employment alternatives available to them; for example, females were a ready source of supply in nursing but now have more career opportunities.

Institutional barriers, such as registration requirements also increase the time taken for quantity to adjust in response to increase in demand. For example, the power of doctors controlling registration and training was seen to restrict the numbers in training in the United States (Seldon, Jung and Cavazos 1998).

Many of the reforms to the training system in Australia of the last 15 years have been directed at improving the responsiveness to the training market. This has included making a proportion of public funds contestable, especially in relation to the delivery of apprenticeships and traineeships, among public and private providers. It has also meant changes in the content of courses to reflect the workplace requirements more closely.

To insure against leakage of potential workers to other occupations or employers, some employers offer bonded contracts to students in training. For a long period up to the early 1970s, state governments in Australia provided financial support for fees and living expenses to student teachers who in return they promised to teach in government schools for some minimum period.

Apparent shortages may also arise because some employers prefer employees from certain demographic or ethnic groups (for example, the preference for employing young people in the ICT industry in the 1990s). They can also arise from the two-year experience syndrome among some employers who are reluctant to hire employees without experience.

Workers leaving an occupation are a factor reducing supply. The rate of turnover of workers varies by age and sex and across occupational groups. The net rate--the number leaving an occupation net of re-entrants to it--tends to be lower in the professional and more skilled occupations than in lower skilled occupations (Shah and Burke 2003).

Net immigration affects supply, too, and can vary over time with changes in migration policies and the relative opportunities in Australia and overseas. The government has more control over immigration and less over emigration. The time lag from the moment a particular skills shortage is identified to when the first migrants with those skills arrive can, however, be long.

Lack of labour market information

The availability of good quality labour market information to firms, households and education and curriculum planners is the life-blood of an efficient market. Lack of reliable labour market information can hamper the speed of the market adjustment process and therefore the duration of shortages. The time taken for information to flow to workers about new opportunities and the time taken by them to take advantage of those opportunities add to the time lag. Since some workers may need training, inefficiencies arise when there is lack of transparency in the links between education and training courses and labour market opportunities (Heijke and Borghans 1998).

Identifying and measuring skills shortages

To develop policy responses to skills imbalance between supply and demand, it is necessary to have an indication of its size, its causes, the extent to which it affects production and whether, on current trends and policy settings, it will diminish rapidly. It is rarely possible to provide an unambiguous single measure of skills imbalance. It is thus important to use a range of measures. Approaches to identifying shortages and estimating their extent are varied. In general, the measures fall in two broad classes (Roy, Henson and Lavoie 1996).

The first includes market economic indicators, such as vacancy, hiring and separation rates, relative wage movements and employment and unemployment changes to infer imbalances for particular occupation groups. These measures provide a market-wide perspective and identify shortages only if there is an insufficient number of appropriately skilled people in the market to fill vacancies at the going wage rates and employment conditions.

Slowness in the adjustment of supply

The speed at which supply adjusts depends on a number of factors. The main factors are the length of training required and the responses of the training institutions to changes in demand. Quite often education and training requirements take a long time to complete, although refresher courses for those who are returning to the same occupation after an extended period of absence from it can accelerate the process. Supply might vary because of changes in education and training provision and changes in preferences for various forms of work and to demographic changes including ageing, emigration and immigration.

Educational institutions adjust only slowly in offering places for particular courses. The time taken for training authorities to recognise a particular need, the internal budgetary processes in public vocational education and training and higher education institutions and the current staffing patterns mean that the number of places provided in areas that are in shortage adjust slowly over time.

The expansion of publicly funded places is not the only factor affecting the numbers. Students must be attracted to take up the new places, and in the case of apprenticeships secure contracts with employers. A decrease in student demand for places can occur when individuals traditionally trained in a given area have other more attractive employment alternatives available to them; for example, females were a ready source of supply in nursing but now have more career opportunities.

Institutional barriers, such as registration requirements also increase the time taken for quantity to adjust in response to increase in demand. For example, the power of doctors controlling registration and training was seen to restrict the numbers in training in the United States (Seldon, Jung and Cavazos 1998).

Many of the reforms to the training system in Australia of the last 15 years have been directed at improving the responsiveness to the training market. This has included making a proportion of public funds contestable, especially in relation to the delivery of apprenticeships and traineeships, among public and private providers. It has also meant changes in the content of courses to reflect the workplace requirements more closely.

To insure against leakage of potential workers to other occupations or employers, some employers offer bonded contracts to students in training. For a long period up to the early 1970s, state governments in Australia provided financial support for fees and living expenses to student teachers who in return they promised to teach in government schools for some minimum period.

Apparent shortages may also arise because some employers prefer employees from certain demographic or ethnic groups (for example, the preference for particular skill may oscillate from surplus to shortage, but over the period of a business cycle, no imbalance may be evident. Similarly, in occupations where the demand is very seasonal, analysis by seasons will show a shortage or a surplus, but if it is over a year then no imbalance may be evident.

The characteristic of the occupational group should guide the choice of the period for analysis of imbalances. The choice also depends on whether one is interested in the short-run fluctuations or the long-run fundamental structural imbalances. The analysis should be in the context of the general economic cycle within which the data are situated.

Spatial dimension

The performance of a labour market varies across regions. One aspect of this geographical disparity could be that while some regions have shortages of particular skills others have surpluses. The spatial imbalances are more persistent in some countries than in others because of the differences in wage flexibility and adjustment costs in terms of geographic mobility (OECD 1994).

Inclusion of spatial dimension to the analysis is important, as internal migration sometimes becomes part of the solution to the problem of skills imbalances. Its inclusion in the analysis, however, should consider the rate of internal migration within the occupation under investigation. Since financial costs of moving relative to expected earnings are lower for a higher skilled worker than a lower skilled one, and training costs increase with skill level, relocation would be a more viable solution for higher skilled workers. Hence, one would expect a higher rate of internal mobility with a national scope in higher skilled occupations, and a low rate with a regional or local scope in lower skilled occupations (Roy, Henson and Lavoie 1996).

Spatial analysis of imbalances adds further pressure on the detail required in the data, especially if it is to be at a disaggregated level. Thus the quality of the analysis will very much depend on the quality of the data.

Indicators of skills shortages

Since a single indicator of skills shortage does not exist, a suite of indicators is used to assess job market conditions. Based on these indicators and judgment, a view is then formed on the likelihood of a shortage or a surplus. Often these indicators have to be observed over time before firm conclusions can be drawn on the existence of an imbalance. A number of the indicators are discussed below.

Hard to fill vacancies and vacancy rates

Unfilled vacancy rates are sometimes used to assess the tightness in the labour market. A large number of unfilled vacancies that are hard-to-fill are indicative of a skills shortage. As discussed below, DEWR has developed a program for estimating hard-to-fill vacancies and recruitment difficulties as a basis for its national skills shortage list.

There are, however, problems in interpreting unfilled vacancy statistics from ad-hoc employer-based surveys. These problems are mainly to do with the inconsistent interpretation of various aspects of skills shortages by employers in self-reported surveys (Green, Machin and Wilkinson 1998). For example, of the 560 000 vacancies estimated in England from the Employers Skills Survey in 1999, 255 000 were hard-to-fill, but on closer analysis only 110 000 of these were 'genuine' market shortages; the rest were due to other recruitment difficulties (National Skills Task Force 2000).

A properly constructed and regularly conducted interview-based survey in which employers are presented with precise concepts of skills shortages may overcome some of the problems. Such surveys are costly and complicated to design. Finally, because of job-to-job turnover of workers within and across occupations, vacancy statistics tend to overestimate market needs and therefore the 'normal' vacancy rates for various occupational groups have to be determined.

Unemployment rates

Occupational unemployment rates can provide indications of skills imbalances-a high rate indicating a surplus and a low rate a shortage. However, it would be naive to think that any positive unemployment rate is an indication of a surplus. There are numerous reasons why the observed and equilibrium occupational unemployment rates are never zero. The challenge with this type of analysis is in determining what is the normal (equilibrium) unemployment rate. With currently available data, it is very difficult, at the occupational level, to determine this normal rate.

The usual unemployment data record an unemployed person's occupation in his or her last job. This is not always the occupation that a person is qualified for or the occupation that a person is actively looking for work in. The unemployed include some who are voluntarily unemployed and others who may not be qualified to work in the occupation, so that their numbers overestimate the supply. It is also possible that the number of unemployed under-represent supply. For example, a person who is qualified to work in one occupation, but is temporarily employed in another, is not counted.

Net vacancies

If both vacancy and unemployment statistics are available, then the unemployment-vacancy (UV) technique can be used to assess skills shortages (Roy, Henson and Lavoie 1996). A simple interpretation would be that if the number of vacancies is larger (smaller) than the number of unemployed persons then a shortage (surplus) could be inferred.

A number of assumptions underlie this technique. First, it has to be assumed that vacancies and the unemployed represent unmet demand and supply under current market conditions. This assumption does not always hold because, as discussed above, the unemployment rate can somewhat exaggerate supply. Therefore, if vacancies reflect true unmet demand, the UV analysis is likely to bias results toward a shortage.

The UV analysis is often used to study the nexus between unemployment and vacancies at the aggregate level. At any point in time, unemployment and vacancies may co-exist largely because of labour turnover, and not necessarily because of deficiencies in the job matching process. The UV relationship is sensitive to the economic cycle, with large numbers of unfilled vacancies during periods of high economic activity and high unemployment during periods of economic decline. Again, while this analysis is possible at an aggregate level, it can prove to be difficult at the occupational level.

Wages

In a market economy, where wages and prices moved freely, occupational labour imbalance would translate to changes in relative labour remuneration. A shortage should pull relative remuneration up while a surplus should push it down. Changes in relative wages, taken to indicate total remuneration, have therefore been used to study occupational labour market imbalances. As the market slowly adjusts to the shortage for particular skills by increasing wages, this trend will be observed as increasing wage differential over time.

There are, however, practical problems in interpreting the wage differential data, because a proper analysis requires knowing the supply and demand elasticities, that is, how much the quantities supplied and demanded change in response to a change in pay. For example, if the supply and demand is highly inelastic (large change in wages result in a small change in supply) then wage-based estimates of imbalances will be higher than quantity-based estimates. The opposite will be true if the supply and demand is highly elastic (a small change in wages results in a large change in supply), and in this case, other things being equal, employers are more likely to be complaining about shortages.

Another practical problem with interpreting wages data arises when institutional factors rather than increased demand drive wages up. For example, anti-discrimination or minimum wage legislation can drive up wages in some occupational groups. Real wages could also reflect the power of the union representing the workers rather than a skills shortage in the occupation. A further difficulty is related to the lack of reliable data for recently hired workers. Wage information on the newly hired is preferred because the normally available average or median occupational wage data can vary because of quality or quantity shifts within the occupational group.

The OECD (1994) provides cross-country evidence suggesting wages are relatively inflexible and do not respond strongly to current market conditions. It contends that this is the main reason for wide dispersion in unemployment rates across occupational groups. Thus with an inflexible wage structure, shocks will result in persistent imbalances which tend to clear through inter-occupational mobility rather than wage adjustments (Roy, Henson and Lavoie 1996).

Other labour market indicators

Other indicators have proved useful in analysing skills imbalances. Their potential usefulness once again depends on data reliability and its careful interpretation. Some of these indicators are hours and intensity of work (e.g. overtime); production levels; change in employment levels; flows of new entrants and leavers; training expenditure by firms; levels of subcontracting; hiring standards; and levels of immigration and emigration.

Employer-based surveys

Employer-based surveys are commonly used to assess skills shortages but hardly ever to assess surpluses. They very much rely on employers' perceptions of skills shortages. Laslett (1992), based on his wide experience in assessing engineering skills in the UK, concluded that skill shortages identified from ad-hoc employer surveys often disappeared when the problem was followed up. Furthermore, he concluded that it is always possible to elicit positive responses to questions about skills shortages.

Investigating the employer perceptions of skills shortages in England, Green, Machin and Wilkinson (1998) found that employers' experience of skills shortages only partially overlapped with the experience of hard-to-fill vacancy or skills gaps. They concluded that the term 'skills shortage' means, to employers, something wider or different from hard-to-fill vacancies. This was not because employers cannot interpret questions on skills shortage, but because their interpretations were not uniform. In addition, a number of employers seemed to give quite inconsistent responses. This study underscores the importance of careful design, implementation and interpretation of any employer-based survey of shortages. A well-designed establishment-based survey within the framework of an economic indicators approach, in contrast to an ad-hoc collection of employers' perspectives on shortages, can however yield useful information.

Structural models

Structural models of firms adjusting to internal labour market imbalance and of the economy-wide effects of supply and demand shocks have been used to assess current and future labour market shortages (Wilson 1992; Hughes 1994; Heijke 1994; Boothby, goth and Roy 1995; Adams et al. 1994). There are pros and cons in using these models.

One advantage of structural models is that they provide some initial baseline data for more rigorous study of imbalances. There are, however, drawbacks limiting their use in practice. First, in general, the models have very large requirements of quality data to calibrate and maintain them. Consequently, many of these models are highly aggregated and simplified. Usually, the more disaggregated the models are, the simpler is their specification. Second, most models project occupational demand requirements rather than current or future market imbalances, which limits their usefulness unless ancillary information from other sources is available on turnover and supply.

In-depth studies

The above discussion suggests that finding evidence of apparent skills shortages is relatively easy but that assessing genuine market shortages is more difficult. A holistic approach that considers a raft of measures for a group of inter-related occupations is more likely to provide robust information on imbalances. Roy, Henson and Lavoie (1996) call this approach an in-depth study.

An in-depth study would ideally provide a framework within which to situate the labour market under investigation together with its structure in terms of its demographic, educational and jobs characteristics. In such an approach, not only all available economic indicators, including those on supply, are gathered in a coherent framework in both time series and cross section modes, but also new data are generated through additional surveys and through meetings of the relevant partners who have a stake in the issues. The study would ascertain whether the imbalances had a geographic dimension or not. The analyses would provide indications of how and how fast the occupational market resolves imbalances. Where possible an in-depth study would also include forecasts of growth and various types of turnover and job openings.

In-depth studies are costly if they are done at an economy-wide level. Since they take time to do, it may only be feasible to use the technique to investigate markets in which imbalances are persistent and adjustment periods are long.

Studies of skills shortages in Australia

This section, first, briefly reviews DEWR's method of compiling the national skills shortage list. Secondly, it illustrates, using the example of the ICT industry, the difficulties of assessing skills shortages in a dynamic sector and also the problems that can arise when policy is developed without reliable data and/or thorough analysis.

DEWR's national skills shortage list

In Australia, DEWR annually publishes a list of professional and trade occupations in which skills shortages have been assessed (DEWR 2004a). The list, together with other information, is used to produce the Migration Occupations in Demand List (MODL) for prioritising visa applications under the Skills Migration program. This is the only work that monitors and collects data on national and regional skills shortages and recruitment difficulties on a regular basis in Australia. The program covers assessment of skills shortages in trades, professionals and information and communication technology (ICT) occupations. The assessment is carded out on an annual basis for trades and professionals and on a six-monthly basis for ICT skills.

To assess shortages, a telephone-based survey targeting employers who have recently advertised for positions in selected occupations, mainly in the major metropolitan (7) newspapers, is conducted through DEWR state offices. Employer contact is continued over the following weeks to find whether the vacancy is filled or not. Hard-to-fill vacancies are those that are still unfilled four weeks (six weeks for professional occupations) after initial advertising. The state-level information is collated at the national office to derive the imbalance situation at the national level.

Under the DEWR definition, shortages are typically for 'specialised and experienced workers' and can co-exist with relatively high unemployment overall in the occupation. Thus, an occupation can be assessed as being in shortage even though not all specialisations are in shortage. Furthermore, occupations may be in shortage in a particular geographical area and not in others. If shortage is evident in the three largest states or in a majority of states for a particular specialisation in an occupation, then the shortage is declared nation-wide.

The methodology has two advantages over the usual ad-hoc employer-based surveys. Firstly, only employers who have actually advertised positions are included in the survey. This means that 'opinions' of employers who are not in the process of recruiting are not used for assessing shortages. Secondly, a follow-up survey ascertains whether or not vacancies remain unfilled and inquires about reasons.

Skills shortages in ICT

At the height of the dotcom bubble in late 1990s, the National Office of the Information Economy (NOIE) reported on two studies on skill shortages in the ICT industry (NOIE 1998; IT&T Task Force 1999 (8)). The second report projected a requirement of 180,000 additional workers, 90,000 with higher education qualifications, in ICT occupations for 1999-2004. The projections were based on data from a survey of employers carried out by the consulting firm Deloitte Touche Tohmatsu. If these projections were accurate, then even with the most optimistic assumptions about the output of graduates with these skills from training institutions, it seemed that severe skills shortages were unavoidable unless other policies were quickly implemented.

Three more studies followed and their reports published in 2001. The first was put out by the Business Council of Australia (BCA). Assuming graduate completions of about 40,000 over five years, the study used the projections from IT&T Task Force (1999) to estimate unmet demand of 50,000 for higher education IT graduates to 2004 (Boston Consulting Group with assistance from Arthur Andersen Freehills 2001). The second study, Skilling People for an Information Society by the New South Wales Government, also quoting figures from IT&T Skills Task Force (1999), reiterated the impending critical shortages (ICT Industry Skills Consultative Group 2001). The third study was commissioned by the Australian Information Industry Association. It reported a shortfall of 27,500 ICT workers to 2004 (Centre for International Economics 2001). The figure was arrived at by starting with the baseline figure of 50,000 from the BCA report. Then adjustments were made for the impact of the softening in the demand for ICT workers, increased enrolment in IT courses and government policy changes with respect to visa requirements for permanent residency for overseas students who completed their courses in Australia (Commonwealth of Australia 2001).

A further study for NOIE by Allen Consulting in 2002 continued to report a shortage of labour with the necessary IT skills, with employers citing this as a barrier to more involvement in the information economy (Allen Consulting Group 2002).

By early 2003, the dotcom bubble had truly burst and DEWR reported only some specialisations in ICT in the national skills shortage list. The Victorian government published further statistics indicating the softening of the ICT labour market in Victoria in the second half of 2002. These included a drop of 8 per cent in the average salary for ICT personnel; a 41 per cent fall in the number of advertisements for jobs; and job applicants with much higher skills applying for jobs requiring lower skills (Department of Infrastructure 2003).

Finally, it is useful to note a couple of media reports on the subject. In March 2003, Graeme Philipson wrote an article titled: 'Skills shortage is a wretched myth' in The Age. He wrote:

   Unemployment in the IT industry is at an all-time high, according to
   ... Australian Computer Society (ACS) ... Tony Abbot, the Minister
   for Employment and Workplace Relations, says 'ICT skills shortages
   have eased considerably ... there was a rise of 14,400 in the
   employment of ICT professionals in the year to 2002' ... Department
   of Immigration is still granting temporary visas to ICT workers
   stating 'ICT vacancy figure is far greater than any other
   occupation' ... Alston (the Minister for Communication and
   Information Technology) dismissed the ACS report as 'blatant
   grandstanding'.(Philipson 2003).

A month later ICTRA reported 14,360 'real' job ads on their website, an increase of 32 per cent over the previous three months (ICTRA 2003). In September 2003, Chelsey Martin in an article titled: 'Fear of skills shortage as students shun IT' in the Australian Financial Review, reported a drop of 25 per cent in the enrolments for IT programs in universities in the last three years (Martin 2003).

The above illustrates the difficulty in assessing skills shortages, especially in a dynamic industry such as ICT. It also shows the lack of critical evaluation of the projections reported in IT&T Task Force (1999) by most subsequent studies, which used them as the starting point in their own evaluation even though the labour market for ICT skills had already begun to soften considerably by the time the studies were written. Furthermore, it illustrates the danger of relying too heavily on one employer survey to gauge the level of future skills shortages. The dearth of documentation on the methodology for the survey prevents comment on its reliability. Employer-based surveys provide only one dimension of a multidimensional matrix of information required to assess skills shortages.

Public policy responses to shortages

In a competitive market, there are always going to be unfilled vacancies, and a proportion of these could be considered hard-to-fill vacancies. When the number of unfilled vacancies reaches some 'intolerable' level, there are calls for intervention in the market. In some cases, however, the observed imbalance may be socially optimal for the particular labour market, and public intervention would simply be adding to costs of adjustment that are being borne by the market (Roy, Henson and Lavoie 1996). When, therefore, is it appropriate for government to intervene? Equally important are the questions of why and what of government interventions.

Labour markets may be inefficient due to a range of factors--externalities, legislative or institutional. Externalities occur when an activity (e.g. training) undertaken by individuals or firms leads to benefits (or costs) to other individuals or firms. If the externalities are significant, then the result could be a low level of training activity and a likelihood of skills shortages or gaps. Such things as wages policy, subsidies, quotas on registration and managed supply or demand in some government employment can reduce the market-clearing role of prices and wages. Other factors that could contribute to the inefficiency in the market are low labour mobility and a lack of good quality market information.

Once market imbalances, including their severity and causes, have been identified, the next question arises as to what government action, in terms of its effectiveness and cost, is to be taken. The course of action will depend on the specific labour market that is in imbalance, as each market may need a tailored solution. This will involve consideration of the magnitude and severity of the shortage; evidence of cyclical factors; evidence of long-term persistence; the likely extent of substitution with other types of labour; the geographical distribution of the shortage; and the specific types of skills that are in short supply.

Information on skills shortages, recruitment difficulties and skills gaps provides some of the data that are needed to inform public policy-making on education and training. As considered in this paper, shortages are caused by a number of factors affecting demand and supply, and there may be a range of ways whereby they can be brought closer to balance. The rest of this section discusses some of the options available for public policy to deal with current skill shortages and to avoid future ones.

Publicly provided training

Provision of publicly funded training places is one of the ways of overcoming skills shortages and gaps. Public provision is usually rationalised in terms of efficiency and equity arguments. Since economic performance is negatively affected by skills gaps, it is argued that more training will result in an adaptable workforce. The training has, however, to be appropriate to be effective.

Publicly funded training to satisfy employer needs requires a thorough and extensive knowledge of the workings of the particular labour market because not all such needs are most effectively satisfied by these means. While initially it may appear that government action is called for to do something about the shortage, it may not always be the best first option. Often the market will take sufficient corrective actions to alleviate shortages. Government intervention may be counterproductive if its effect is too late or it over-corrects the problem. The recent inquiry into future skills in South Australia concluded that 'government needs to resist the urgings and the urge to do something' skills shortages (Government of South Australia 2003). The inquiry found virtually no evidence to suggest that skills shortages, where they exist, are the result of an inadequate training system. The submission of the Construction Industry Training Board (South Australia) to the inquiry included:

   ... skill is not a commodity that can be readily stockpiled for use
   when demand for it emerges ... Much is made of current or impending
   skill shortages. Often it is contended that this reflects a failure
   on the part of the 'training system'. In fact, rarely is the
   training system to blame.

The nature of the intervention will depend on whether the problem is a skills shortage, recruitment difficulties or a skills gap. Recruitment difficulties may not need publicly provided training as a remedy. Raising employers' awareness of the causes of recruitment difficulties and possible solutions for them are important issues.

If we do conclude that there is a need for a training response, there are still important questions about its cost, how it should be financed, how it will be provided and how students or workers will be attracted to it. The policy choices need to be seen in the context of a commitment to lift the overall skill level of the workforce. It should also be remembered that the public system has a major role in the promotion of equity and education for citizenship and in personal development.

The matters that can be considered in relation to shortages include:

* expansion of the available places in the areas of shortage, subject to cost considerations;

* incentives to enrolment (or provision of contracts of training) and the places made available;

* development of accelerated courses in the skill areas in most demand;

* ensuring that underpinning knowledge and generic skills are developed to allow quicker retraining; and

* development of new courses of training in emerging areas.

Encouraging training in industry

Employers' capacity to respond to shortages will be greater, the greater is their awareness of the importance of training and the greater is their involvement in planning and provision. Engendering a training culture among employers takes time, as it requires attitudinal changes. Considerable attention has been given to this by training authorities in Australia in the last 15 years. Improving the training culture requires employers to take a long-term perspective on training and its benefits for innovation and profits. Employers want to recoup returns on investment in training and are naturally concerned about undertaking the costs of training where there is the danger that much of the benefit may be gained by other firms, particularly in dynamic markets where margins are small (Government of South Australia 2003).

One way of increasing employer involvement and responsibility for training, tried by various governments, has been training levies. Levies have a longer history in European countries than in Australia. The Training Guarantee Levy was tried in Australia in the late 1980s, but abandoned in 1994 because of employer objections. It is a matter of conjecture whether the levy would have succeeded given fine-tuning and more time.

It has been suggested that a training-levy system that is owned and controlled by industry groups, including unions, may have a better chance of success than a system that is national (Government of South Australia 2003; Gasskov 2002). Any training-levy system has to ensure equity of access to training for all workers, particularly those who are least qualified and who have the least bargaining power.

Employer training can also be encouraged by tax relief or subsidies to firms commensurate with the amount of training they provide their employees. The main example in Australia is the subsidies for the employment of apprentices and trainees.

Labour market information

Transparency is essential for the efficient functioning of a market. Pertinent, reliable, up-to-date and generally available information about various aspects of the market provides this transparency and helps clear imbalances in the market. The information must relate not just to a particular market but to all markets. It must be available on current wage rates, rates of return, unemployment rates, job openings and supplies of workers in different sub-markets.

Producing quality information is expensive. If this function is left to the free market, the provision is likely to be less than what would be regarded as socially adequate because of the public good nature of the information. Hence, governments have an important role to play in providing it.

DEWR produces much web-based information on careers, occupational certification, qualifications and job prospects. It is important to emphasise the importance of replacement needs or turnover information, because often job openings due to turnover are more numerous than job openings due to growth, if indeed there is growth in the occupation. Even when there is contraction in an occupation job openings may exist because of turnover.

It is difficult to determine what amount of information governments should provide, became after a certain point there are diminishing returns to increasing investment.

Immigration

Immigration can be used as a policy instrument to overcome skilled and unskilled labour shortages. Countries such as Australia and Canada have skilled migration programs to meet expected shortfalls in particular types of skills. These programs attempt to identify skills in shortage and potential immigrants with these skills are given priority over others. In general, the programs are not open-ended, annual quotas put a ceiling to the number who are allowed in, and occupations in the priority list change. One of the problems with this system is that the arrival of migrants, with appropriate skills, often lags by a considerable time the actual occurrence of the shortage. Sometimes the market has already corrected itself by the time the migrants arrive. The migrants then find themselves having to work in alternative, and sometimes lower skill, occupations. Another problem is that migrants generally prefer settling in large metropolitan areas while the shortages are often in regional or remote areas.

The net flow of skilled labour is usually from developing to developed countries. Notwithstanding the right of people to realise their own potential, economically and personally, ethical issues do arise when aggressive 'pull factors' result in an exodus of scarce skilled personnel from a developing country which can least afford to lose them. In general, unless there is a bilateral government-to-government agreement, the country of origin has no recourse to recoup the investment it has made in training the individuals who leave. Some remittances do flow back from the migrant workers, but they are hardly enough compensation for what is being lost.

Encouraging mobility

Assistance to encourage mobility may help alleviate geographical skills imbalances, but an important issue is that qualifications be recognised throughout Australia. The Australian Quality Training Framework has been developed to provide a set of nationally agreed standards to ensure that all qualifications issued by Registered Training Organisations are recognised throughout the country.

Internal migration programs are more likely to succeed if they are backed with comprehensive information systems, not only about the labour market but also about other services, such as schooling, health and homing. Large disparities in homing costs between different regions can inhibit worker flows, at least in one direction. Whether governments should provide financial assistance for internal migration or not is a difficult policy issue and depends on how strategic the problem is.

Concluding remarks

This paper has provided a review of the concepts, measurement issues about and responses to skills shortages. Economic theory suggests that skills imbalances are a permanent feature of a competitive market. Without major hindrances to the market adjustment process, and without externalities, these imbalances resolve over time and do not require intervention. There will always be unfilled vacancies co-existing with unemployment because of continual adjustment of the market towards equilibrium. Adjustment usually takes time because of the speed at which the information about the changes in the market is dissipated and the time it takes an individual to acquire new skills.

Real-world labour markets, however, are seldom perfect. Government action to correct the market imperfection may be justified if there is strong evidence of sub-optimality. Before any effective policy is designed and implemented, it is important to understand the operation of the current market and to investigate the reasons for its failure. Lack of good empirical knowledge of how markets adjust, and the time lag from when an imbalance is observed to the impact of a policy makes effective intervention difficult in practice.

Care needs to be exercised in interpreting ad-hoc employer-based surveys on shortages because of the varying meanings employers attach to the germ. The best way to assess skills shortages, it seems, is through in-depth studies. Such studies provide an opportunity to understand the distinctive features of the markets for particular skills. These understandings are essential for identifying and finding solutions.

With rapid technological changes, there will always be skills gaps in the workforce. The rate at which these are bridged depends on the training system in place. If employers rely on the public training system to finance and bridge them, then the adjustment process will be longer because of the bureaucratic necessities and the time it takes third parties to gear up with new curricula and delivery strategies. If the firms were to take greater responsibility to bridge the employee skills gaps, then the adjustment would be much faster and the training could be tailored specific to the firm, thus reducing problems of externalities.

Endnotes

(1) An earlier version of this paper was presented at the 2004 Labour Market Research Workshop in Adelaide. The authors wish to thank Michael Long and Tom Karmel and two anonymous referees for constructive comments on earlier versions of the paper. The initial work for this paper was funded by the Australian National Training Authority (ANTA).

(2) Although Becker discusses the acquisition of such skills through on-the-job training this restriction is not necessary.

(3) In practice firms often finance training which is of value to other employers, in part because neither the workers nor other employers have full-information on the worker's skills (Stem 1995, Long et al 2000).

(4) In the UK the National Skills Task Force (2000) uses the term other recruitment difficulties to describe this situation while the term recruitment difficulty has a broader meaning.

(5) See Roy, Henson and Lavoie (1996) and Veneri (1999) for a review.

(6) In Australia, this classification is the ABS developed Australian Standard Classification of Occupations (ASCO).

(7) Some pilot studies have recently been conducted to extend the analysis to regional areas.

(8) Only the executive summary of the report could be accessed.

References

Adams, P, Dixon, P, McDonald, D, Meagher, G and Parmenter, B (1994), 'Forecasts for the Australian economy using the MONASH model', International Journal of Forecasting, vol. 10, pp. 557-571.

Akerlof, G. A and Yellen, JL (1986), Efficiency Wage Models of the Labour Market, Cambridge University Press.

Allen Consulting Group (2002), Australia's Information Economy: The Big picture, A Report for the National Office of the Information Economy (NOIE), Allen Consulting Group, Melbourne.

Arrow, K. J. and Capron, W. M. (1959), 'Dynamic Shortages and Price Rises: The Engineer-Scientist Case', Quarterly Journal of Economics, vol. 73, pp. 292-308.

Becker, G. (1962), 'Investment in Human Capital: A Theoretical Analysis', Journal of Political Economy, vol. 70 (supplement), pp. 9-49.

Blank, D. and Stigler, G. (1957), The Demand and Supply of Scientific Personnel, National Bureau of Economic Research, New York.

Boothby, D., Roth,. and Roy, R. (1995), 'The Canadian occupational projection system: A methodological enhancement', Research paper No. R-95-2, Human Resources Development Canada, Hull, Quebec.

Borghans, L., de Grip, A. and van Smoorenburg, M. (1998), Measuring Skill Shortages, ROA-R-1998/4E, Research Centre for Education and the Labour Market, Maastricht.

Borland, J. (2002), 'The market for medical specialists in Australia', Proceedings of Health Policy Roundtable, Productivity Commission, Melbourne, pp. 245-292.

Boston Consulting Group (with assistance from Arthur Andersen Freehills) (2001), BCA E-Business Roundtable, Business Council of Australia, Melbourne.

Bosworth, D., Dutton, P. and Lewis, J. (eds.) (1992), Skill Shortages: Causes and Consequences, Avebury, Aldershot.

Centre for International Economics (2001), Breaking the skills barrier: Demonstrating the benefits of investment in ICT higher education in Australia, Report to Australian Information Industry Association Ltd, Centre for International Economics, Canberra.

Commonwealth of Australia (2001), Backing Australia's Ability, Department of Education, Science and Training, viewed 10 Oct 2003, <http://backingaus.innovation.gov.au/Skills/index.htm>.

Department of Employment and Workplace Relations (DEWR) (2004a), National Skills Shortage Lists Australia--2004, DEWR, viewed 1 March 2005, <http://www.workplace.gov.au/WP/Content /Files/WP/EmploymentPublications/NSSTradesupdateDecember2004V2.pdf>.

Department of Employment and Workplace Relations (DEWR) (2004b), Job Outlook, DEWR, viewed 1 March 2005, http://www.workplace.gov.au/WP/Content/Files/WP/EmploymentPublications /JOMain2004.pdf

Department of Infrastructure (2003), ICT Skills Snapshot: The State of ICT Skills in Victoria, Department of Infrastructure, Melbourne.

Friedman, M. and Kuznets, S. (1945), Income from Independent Professional Practice, National Bureau of Economic Research, New York.

Frogner, M. (2002), 'Skills shortages', Labour Market Trends, vol. 110, pp. 17-27.

Gasskov, V. (2002), 'Industry support for training, an international perspective', paper presented to CEET National Conference, Melbourne, 26 August.

Government of South Australia (Reviewer K. Schofield) (2003), Skills for the Future, Final Report of the Ministerial Inquiry, Adelaide.

Green, F, Machin, S. and Wilkinson, D. (1998), 'The meaning and determinants of skills shortages', Oxford Bulletin of Economics and Statistics, vol. 60, pp. 165-187.

Heijke, H. (ed.) (1994), Forecasting the Labour Market by Occupation and Education, Kluwer Academic Publishers, Boston.

Heijke, H. and Borghans, L. (eds.) (1998), Towards a Transparent Labour Market for Educational Decisions, Ashgate, Aldershot.

Hughes, G. (1994), 'An overview of occupational forecasting in OECD countries', International Contributions to Labour Studies, vol. 4, pp. 129-144.

ICT Industry Skills Consultative Group (2001), Skilling People for an Information Society, New South Wales Government, Sydney.

Information Technology Contract and Recruitment Association (ICTRA) (2003), 'Real' jobs grow by 18.5 % on IT2 in March, ICTRA, viewed 5 Sept 2003, <http://www.itcra.com/news.jsp>.

IT & T Task Force (1999), Future demand for IT & T skills in Australia, National Office for the Information Economy, viewed 8/10/2003, <http://www.noie.gov.au/publications/1999.htm>.

Keep, E. and Mahew, K. (1999), 'The assessment: knowledge, skills and competitiveness', Oxford Review of Economic Policy, vol. 15, pp. 1-15.

Laslett, R. (1992), 'Surveys of skill shortages in engineering labour markets', in Bosworth D., Dutton, P. and Lewis, J. (eds), Skill Shortages: Causes and Consequences, Avebury, Aldershot.

Long, M., Ryan, R., Burke, G. and Hopkins, S. (2000), Enterprise-Based Education and Training, Ministry of Education, NZ, Wellington.

Martin, C (2003), 'Fear of skills shortage as students shun IT', Australian Financial Review, September 6.

National Office of the Information Economy (NOIE) (1998), Skill shortages in Australia's IT & T industries, NOIE, viewed 7 Oct 2003, <http://www.noie.gov.au /projects/aceess/Education/discpaper_dec98/skills.html#Shortterm>.

National Skills Task Force (1998), Towards a National Skills Agenda: First Report of the national Skills Task Force, SKT 1, Department for Education and Employment (UK), Nottingham.

National Skills Task Force (2000), Skills for all: Research Report from the National Skills Task Force, SKT 29, Department for Education and Employment (UK), Nottingham.

OECD (1994), OECD Jobs Study: Evidence and Explanations, OECD, Paris.

Philipson, G. (2003), 'Skills shortage is a wretched myth', The Age, 4 March, viewed 10 Sept 2003, <http:/www.theage.com.au/articles/2003/03/03/1046540116777.html>.

Rainbird, H. (1992), 'Union perspectives on skill shortages', in Bosworth, D., Dutton, P. and Lewis, J. (eds.) Skill Shortages: Causes and Consequences, Avebury, Aldershot.

Roy, R., Henson, H. and Lavoie, C. (1996), A Primer on Skill Shortages in Canada, R-96-8E, Human Resources Development Canada, Hull, Canada.

Seldon, B., Jung, C. and Cavazos, R. (1998), 'Market power among physicians in the U.S., 1983-1991', Quarterly Review of Economics and Finance, vol. 38, pp. 799-824.

Shah, C. and Burke, G. (2003), Job turnover: replacement needs and vacancies by occupation, Report for the Department of Employment and Workplace Relations, Canberra.

Solow, R. M. (1979), 'Another Possible Source of Wage Stickiness', Journal of Macroeconomics, vol. 1, 79-82.

Stern, D 1995, 'Market failure in firm-based education and training', in Carnoy M. (ed.), International Encyclopaedia of Economics of Education (2nd edn.), Elsevier Science, Oxford.

Veneri, C. (1999), 'Can occupational labor shortages be identified using available data?', Monthly Labor Review, March, pp. 15-21.

Wilson, R. (1992), 'Forecasting occupational change and skill imbalances', in Bosworth D., Dutton, P. and Lewis, J. (eds), Skill Shortages: Causes and Consequences, Avebury, Aldershot.

Chandra Shah and Gerald Burke, Monash University-ACER Centre for Economics of Education and Training