Impact of flexible labour market arrangements in the machinery electrical and electronic industries

Report for discussion at the Tripartite Meeting on
the Impact of Flexible labour market Arrangements
in theMachinery, Electrical and Electronic Industries

Part 7    

Copyright ® 1999 International Labour Organization (ILO)

7. Labour compensation

In advanced manufacturing, labour costs represent a relatively small part of total costs. Moreover, there is much scope for productivity improvements. Management has discovered the productive role of shop-floor experience and of workers' skills and education. The main issue therefore is how to design compensation systems that provide appropriate incentives rather than minimization of labour costs as such. Typical features of new compensation systems will be presented first. But they must be set against broader developments that took place in recent years and which affect the level and distribution of earnings. These trends are mirrored in the machinery industry, as case-studies illustrate. They are presented after an analysis of labour costs in the machinery industry, gender wage differentials, and some aspects of taxation and social security contributions. This chapter focuses on high-income OECD countries, for which more detailed data are available. Other countries were examined in Chapter 2, on the basis of UNIDO data; more will be said in Chapter 9 when discussing the special case of EPZs.

7.1. Some characteristics of new
compensation systems

New and flexible forms of work organization are associated with revised compensation systems that provide appropriate incentives. Some of the pay systems in use today were developed 40 or 50 years ago. Often they only provided for remuneration of a specific job within a rigid and hierarchical division of labour. They are now proving to be an obstacle for the introduction of more flexible work arrangements. In the most up-to-date companies traditional pay differentials are no longer workable: between blue and white-collar workers, men and women, full-time and part-time employees, time and piece-rate workers, or members of the same team. New payment systems typically rest upon broad job descriptions, a corresponding reduction in the number of pay grades, incentives to acquire extra qualifications, higher valuation of new job requirements, supplements to basic wages related to firm performance or continuous improvement, and internal job equity.⁽¹⁾

A survey of 25 British engineering firms found that 16 had introduced single-status employment policies (often associated with Japanese working practices), with a single grade on the shop floor, in order to break down barriers, engender a feeling of equality, and thus improve morale and flexibility; 13 schemes had been introduced in the past five years. Holidays, sick pay and canteen facilities had been harmonized in all 16 cases; the next most commonly harmonized items were pension schemes, method of payment, time clocking and private health schemes.⁽²⁾ A 1994 survey of 33 electronics firms in the United Kingdom indicated that all but one operated single-status policies. Of the eight most commonly harmonized arrangements, those most frequently reported were canteen facilities (33 companies), holiday entitlement (31), sick pay (30), pensions (30), working hours (28), time clocking (24), clothing (22) and private health care (13).⁽³⁾

Companies are also shifting to systems that assess employees' current skills, future potential and willingness to learn. They concentrate on employees' development, not past performance; they are "development-led, not measurement-led".⁽⁴⁾ Thus, new criteria have been introduced to gauge the individual component, such as creativity, flexibility, qualifications, ability to work in a team and commitment. But employees and supervisors must be trained for a meaningful application of such evaluation systems.

Moreover, new compensation systems usually include an individual and a group component, mirroring the shift to work teams that assume more responsibilities than in the past. Association of individual workers with the work groups' and/or firms' performance has become an important avenue for promoting entrepreneurship, cooperation, raising productivity and quality, and shifting the focus to customers' needs. However, individual employees have little direct impact on a firm's overall performance. Therefore, specific targets must be selected which are suited to the different work groups and functions within the firm. Preferably, groups are set up as separate cost centres for monitoring purposes, which implies that appropriate and current statistics must be produced to gauge the performance of groups/functions. Cost centres, in contrast to profit centres, offer the advantage of being applicable to non-profitable functions. This system therefore poses fewer problems of internal cooperation among various units than when groups are set up as profit centres. Finally, employees must be kept informed of the firm's overall performance, and they need some basic knowledge of business management to gauge the firm's performance and calculate costs.⁽⁵⁾

Unions play an important role in the planning and implementation of change. For instance, the above-mentioned survey of 25 British engineering firms indicated that in most cases they had been receptive to change and often played an important role in communicating the need for these innovations to the workforce. Unions were unreceptive in only four out of the 22 cases in which negotiations over change took place.⁽⁶⁾ The shift to collective pay premiums related to group/firm performance, however, has one major implication for employees, namely that pay becomes variable and, to some extent, unpredictable; employees share in the entrepreneurial risk of business. The introduction of these premiums may therefore raise difficulties with both staff and unions. Moreover, merit pay and other individual components may lead to greater wage differentials under new pay systems than job classifications and pay grades suggest. Also, since individual benefit components are usually not known to other workers, these systems may in fact reduce transparency and result in a "feeling of uneasiness" among staff rather than engendering a "feeling of equality". Thus, in the Italian machinery industry, for instance, the most common problem at issue in collective agreements was remuneration. Wage claims mainly concerned non-traditional subjects such as the corporate parameters used for the calculation of merit pay increments (productivity, profitability, quality, efficiency and presence).⁽⁷⁾

In French satellite establishments of car assembly plants, new remuneration systems are introduced on an experimental basis in these new sites before attempting extension to the rest of the firm. But this transfer to older plants is often difficult, if not impossible. Unions are not present in new sites, which exhibit a radically different work culture. Thus, one firm reports being unable to reach an agreement with unions over a collective bonus, linked to the performance of the plant, because older staff in the other plants are not willing to "work for others". They are very attached to individual work premiums and to the "working class culture". In one plant, bonuses consisted of 60 per cent individual premium pay and 40 per cent collective component, but the objective was to have only collective premiums in three to four years.

A main drawback of the compensation system used in these new French plants is the substantial (and increasing) gap between what is required of staff and what is remunerated. Only qualifications acquired in the firm are compensated. Thus, production workers start at the bottom of the pay ladder, at or near the minimum wage (SMIC), although firms are becoming increasingly selective in their recruitment. But the aptitudes sought by firms are not part of recognized (and therefore remunerated) qualifications. Moreover, tasks formerly performed by non-production workers, and therefore considered as skilled jobs, are no longer classified as such when they are performed by production workers. Hence, initially, operators derive no advantage with respect to classification or remuneration for the new responsibilities they must assume. The term "polyvalent production agent" is often used in these French plants to describe a job which is generally undervalued. "Strategic jobs" are better paid than others and filled by permanent staff; but replacements are always available. The strategy of multiskilling seeks to train a surplus of persons for each function. The most competent staff are those who can master the greatest number of functions. Competence, not the function, is remunerated; and there is no automatic pay increase with seniority. Thus, to increase their salaries, employees must demand to be trained further. But the pay ladder is short, thus reducing earnings differentials, in keeping with the flatter hierarchical structure of these plants. Finally, each new implantation may lead to a change in the classification and remuneration system of other older plants.⁽⁸⁾

The elimination of job demarcations and the introduction of equal pay systems do not necessarily result in low wages, as the following example illustrates. A Danish machine-building company, which represents cutting-edge practice, took advantage of the introduction of new technology to shift to innovative work and compensation systems. It located the novel high-tech department in a greenfield site, 20 km away from the main factory, in order to break away from the existing labour-management arrangements so as to disengage itself from traditional job demarcations, as well as from a system in which a foreman had a monopoly of planning and scheduling tasks on the shop floor. This role was eliminated in the high-tech department. Unions eventually accepted the elimination of traditional job demarcations between skilled and unskilled jobs and the introduction of an equal pay system in exchange for a considerable wage increase.

Subsequently, a more informal division of labour developed in this new high-tech department, between a relatively responsible operator job on the one hand, and a job role with planning, scheduling, programming and managerial functions on the other. In the new work culture, workers are more involved in issues which are traditionally considered management's areas of responsibility. Only actual knowledge and qualifications, not job demarcations, determine who can carry out which task. Workers are expected to take on considerably more responsibility than in the past, to contribute to the development of products and methods in a creative way and to guarantee a higher productivity level than in the main factory. Management also wishes to extend the relationship of mutual trust. To this effect, it wishes to change the role of shop stewards and integrate them into daily planning, but without formalizing this cooperation, and use them as "ambassadors" for generating a wider understanding among employees of management's initiatives.⁽⁹⁾

7.2. Recent developments

Changes in pay systems in the machinery industry must be set against broader developments, which influence the level and distribution of wages in the sector. The move to flexibility affects all aspects of labour market arrangements. Table 7.1 summarizes some recent government interventions aimed at increasing flexibility in labour and product markets, which have an impact on wage determination; but these measures also highlight the interdependence of various aspects of policy. These developments may account in part for the noticeable wage moderation over the past decade in most OECD countries, compared with the previous recovery in the 1980s. But part of this moderation also seems to be attributable to the decline in trade union density in many countries. Wage moderation did not, however, have the same impact on various groups of workers. Moreover, wage growth has become less sensitive to a rise in unemployment, which would indicate that the equilibrium rate of unemployment has increased.⁽¹⁰⁾

Table 7.1. Recent government interventions to affect wage determination

The relative supply of skilled and less skilled people has shifted. The 1980s saw an upskilling of the labour force in OECD countries. This led to a marked decrease in the share of workers with low educational attainments, while the share of white-collar workers has increased in all countries during the 1970s and 1980s, as was discussed in Chapter 5. The demand for low- skilled workers has also declined, in part because of technological change, but also to some extent as a result of heightened competition from low-wage countries. Thus, although the relative supply of skilled workers has increased, educational premiums have nonetheless risen or remained steady during the 1980s, the increase being strongest in the United States and the United Kingdom. In addition, in a number of cases, a technology wage premium is observable, especially in countries with flexible wages such as the United States, the United Kingdom and Canada. But the question remains to which extent this premium is due to statistically unmeasured skills acquisition such as on-the-job training; firms with advanced technology might be more profitable and therefore able to share profits with employees; finally, such firms might use higher wages as an incentive to raise productivity and attract the best workers.⁽¹¹⁾

The cumulative impact of these developments is that real wages of low-paid workers have actually fallen in Italy, Sweden and the United States; real earnings of women grew faster than those of men in 15 OECD countries, but not in Finland and Sweden; finally, earnings of youth aged 20-24 have fallen relative to prime-age workers.⁽¹²⁾ Thus, wage inequalities increased in many countries in the 1980s, although one would expect the contrary with flatter hierarchies and a reduction in the number of pay grades under new compensation systems. In the early 1990s, however, a rapid increase persisted only in the United Kingdom and the United States. In other countries, the widening gap was either modest, or relatively recent and closely associated with substantial labour market reforms. In a few countries, such as Canada, Finland and most notably Germany, inequalities declined over the last five to ten years.

Wage flexibility (another aspect of flexible labour market arrangements), the shortening of the working week to reduce unemployment and growing wage inequalities raise the question of the incidence of low-paid employment. In general, it appears to be highest in countries with the most pronounced wage inequalities (see table T, The industry in numbers). Thus, in the United States, one-quarter of full-time workers earn less than two-thirds of median earnings (which is used here as threshold for "low-paid" employment), compared to about 7 per cent in Belgium, Finland and Sweden. Low-paid employment is typically heavily concentrated in the wholesale, retail and catering sectors. Youth and women are especially affected. Higher union density and coverage reduce the incidence of low-paid employment, as well as relatively higher minimum wages or more generous unemployment benefits which contribute to raising this wage floor. However, wage floors do not seem to increase unemployment for youth, women and unskilled workers. This suggests that factors other than relative wages would carry greater weight. In addition, there is considerable turnover in low-paid employment. Between 1986 and 1991, low-paid workers who remained in that income group ranged from less than 10 per cent in Denmark to about 30 per cent in the United States. Women are much less likely to move upward than men, except in Finland. Thus, the United States stands out for having high overall wage inequalities but low upward mobility for low-paid workers, in spite of the largest absolute change in individual earnings over 1985-1991 among eight OECD countries (high wage volatility). Entry into and exit out of full-time employment is also much higher for low-paid jobs, a pattern that is especially strong for women.⁽¹³⁾

7.3. The structure of labour costs
in the machinery industry

Non-electrical machinery, compared to electrical machinery, appears to be the subsector with higher real wages, labour's shares in value added and supplements to wages (or non-wage labour costs), in spite of lower labour productivity.⁽¹⁴⁾ This might explain in part the more sizeable relocation of employment to low labour-cost countries that took place in non-electrical machinery and which was noted in Chapter 2. But slower growing real wages (and labour productivity) in non-electrical machinery leads to some convergence between the two subsectors with respect to wages.

Between 1980 and 1994, real wages increased in all countries, except in Mexico in both subsectors, and Luxembourg and the Netherlands in electrical machinery. This growth was most rapid, in electrical machinery, in the Republic of Korea, Norway and Austria; and in non-electrical machinery, in the Republic of Korea, Luxembourg and Norway. In most cases it accelerated in the 1990s, especially in electrical machinery; and in over half the cases real wages expanded more rapidly in electrical than in non-electrical machinery. Hence, some convergence between the two subsectors appears to be taking place with respect to real wages.

Over the same period, labour productivity also increased in all countries except Denmark in non-electrical machinery and Luxembourg in electrical machinery. This growth was more rapid in North America than in Europe in both subsectors. But it was highest, in electrical machinery, in the Republic of Korea, Finland and Japan; and in non-electrical machinery, in the Republic of Korea, Ireland and the United States. Labour productivity generally increased more rapidly in electrical than in non-electrical machinery. Moreover, this growth accelerated in the 1990s in electrical machinery in North America, but otherwise it slowed down, or even became negative, especially in non-electrical machinery. This explains why the increase of wages in non-electrical machinery did not accelerate as much as in electrical machinery. What is a more interesting feature, however, is that declines in labour productivity (in either subsector) were never matched by reductions in wages; the latter kept increasing.

This slowdown of labour productivity relative to wages in non-electrical machinery also shows in the increase of labour's shares over the 1980-94 period. Conversely, in the faster growing electrical machinery subsector, labour's shares have generally declined, because value added increased more rapidly than labour costs.⁽¹⁵⁾ (Labour's share in value added provides a good indicator of the extent to which productivity gains are shared with employees in the form of wage increments.) In 1980, labour's shares in value added were still slightly higher in electrical than in non-electrical machinery in North American and European countries (77-78 per cent on average for the former subsector and 75-76 per cent for the latter). But by 1994, the relative size of these shares was reversed (56-74 per cent for the former subsector and 78-76 per cent for the latter); there was a marked difference in favour of non-electrical machinery. In both subsectors, these changes were more pronounced in North America than in European countries, consistent with the finding above that wages are generally more volatile in North America.

Moreover, in both subsectors, labour's shares always declined after exceeding 80 per cent of value added, except in Sweden in electrical machinery and West Germany in non-electrical machinery. Conversely, when these shares were below 60 per cent, they increased over the period, except in Belgium in non-electrical machinery and in Portugal in electrical machinery. This indicates that convergence is taking place between countries, but not subsectors. In 1992-94, labour's shares were highest, in electrical machinery, in Sweden (101 per cent), Belgium (90 per cent), the United Kingdom (85 per cent) and West Germany (82 per cent); and in non-electrical machinery, in West Germany (89 per cent), Norway (88 per cent), the United Kingdom (84 per cent) and Denmark (83 per cent).⁽¹⁶⁾

Intensified international competition might lead to a convergence of labour costs in world markets. To explore this question table 7.2 presents hourly labour compensation of production workers (converted into current US$ to facilitate international comparisons).⁽¹⁷⁾ Among high-income OECD countries, high rates of growth of real earnings per employee (exceeding 2 per cent per year) are for the most part associated with low labour-cost countries. But convergence is not really taking place, in spite of some catching up of low labour-cost countries. Thus, from 1975 to 1994, the gap (relative to United States wages) widened between the three lowest and the three highest labour-cost countries, when comparing the average labour cost that was calculated for each group; this gap is somewhat larger in electrical than non-electrical machinery.

However, high labour-cost countries do not necessarily have high labour shares in value added. For the present purpose, high labour-cost countries have been defined as those with labour costs exceeding United States levels (i.e. labour-cost indices in table 7.2 are greater than 100, since US = 100). In non-electrical machinery, of the seven high labour-cost countries as defined here, only two had labour shares that exceeded the United States level of 80 per cent. In electrical machinery, however, all ten high labour-cost countries also had labour shares exceeding United States levels, because the United States has a particularly low share in this subsector (56 per cent). In other words, high labour costs can be compensated by a shift to production that has a higher value added content, thus keeping labour shares in value added low.

Among high labour-cost countries, this shift to production with a higher value added content is especially evident in Germany and Denmark. Ratios of "value added intensification" have been calculated as the change in value added over the change in gross output between 1970 and 1995 (see tables Q.1 and Q.3, The industry in numbers). A ratio that exceeds one indicates that value added per unit of production has increased. By 1994, West Germany had the highest labour costs in both subsectors, but it also had the highest ratio of "value added intensification" in non-electrical machinery (1:21) and the third highest in electrical machinery (1:14). The countries that shifted to more value added intensive production are as follows, in decreasing order: in non-electrical machinery, West Germany, Australia, Luxembourg, Denmark, Spain and Ireland; and in electrical machinery, Ireland, Australia, Republic of Korea, West Germany, Spain, Canada, Denmark, the United States and the United Kingdom. However, except for Denmark and West Germany in both subsectors, these are not high labour-cost countries as defined here.

Table 7.2. Labour costs of production workers in the machinery industry
in US$ (1987 SIC codes 35 and 36), 1975-94 (US = 100) ¹

7.4. Gender

In high-income OECD countries, average real wages are usually higher in non-electrical than in electrical machinery, unlike in other countries, as was noted above and in Chapter 2. Relatively lower wages in this subsector, however, might be due to a higher proportion of women working part time, bearing in mind that these figures include both full and part-time employees and that women are more likely than men to work part time. But figures are not available to substantiate this hypothesis. Thus, in the "feminized" subsector of electrical machinery, men and women alike fare worse than in the "masculinized" non-electrical machinery. In relative terms, however, the few data available suggest that overall, women perform somewhat better in the "feminized" subsector, in both OECD and non-OECD countries (see tables 7.3 and 7.4). Gender wage differentials are generally smaller. Moreover, between 1985 and 1996, women's wages as a percentage of men's wages declined less frequently in this subsector than in non-electrical machinery, although in Denmark, Germany, Japan, Sweden, the United Kingdom, Greece and Egypt they declined in both sets of industries.

But this relative advantage remains fragile. In most cases in both subsectors, an increase in the proportion of women was associated with a decline in women's wages relative to those of men, and vice versa. This suggests that the law of supply and demand plays an important role in setting women's relative wages. Moreover, when women's wages already amounted to 80 per cent or more of men's wages they declined more frequently than in the sample of countries as a whole.

Nevertheless, the rate of increase in both sets of industries was most substantial in Turkey and Thailand, although the ratio of women's to men's earnings was already high in non-electrical machinery (88 per cent and 78 per cent, respectively). Conversely, it continued to decline in Japan, where it was already the lowest (less than 50 per cent), except for Malaysia in electrical machinery. Nonetheless, some of the fastest growth occurred in countries where this proportion was relatively low, indicating that overall some "catching up" took place.

These findings, however, obscure the most important elements with respect to the gender wage differential, namely that much of this gap is due to occupational segregation and to women's higher propensity to work part time or on a temporary basis and to take parental leave. Temporary and part-time employment offer poor career prospects, they are often paid less than regular full-time jobs, and they may not provide similar levels of social contributions, thus resulting in insufficient pension entitlements. In the EU, in most countries, inadequate retirement pension is the main reason for poverty; it affects women in particular. Moreover, these forms of employment are still concentrated in low-skilled jobs, thus compounding the problem of low-paid employment.⁽¹⁸⁾ Finally, with the tightening of the labour market, family policy and the reconciliation of work and family have been quietly refashioned into a reduction of the labour supply.⁽¹⁹⁾ Even when it does not lead to withdrawal from the labour force, parental leave frequently has a permanent "scaring effect".

Table 7.3. Women's earnings as a percentage of men's earnings in the
manufacture of machinery, except electrical (ISIC 382, Rev.2),
1985-96 (employees) ¹

Table 7.4. Women's earnings as a percentage of men's earnings in
the manufacture of electrical machinery (ISIC 383, Rev.2),
1985-96 (employees) ¹

7.5. Taxation and social security

Taxation and social security systems are seldom adapted to new flexible work organization, in particular external and numerical flexibility, which raises the question of policy coherence. Social security systems have often been modelled on a lifelong permanent pattern of employment. Benefits are predicated upon a contributory insurance system and eligibility often depends on continuous service. Social security systems will therefore need to be adapted to the various forms of employment people may increasingly experience (such as part-time or "permanently temporary" employment),⁽²⁰⁾ in order to reduce the penalty associated with career breaks, and ensure continuous coverage in spite of changes of employer and/or employment status.

Moreover, the structure of taxation and social security systems may affect employment decisions in undesirable ways. For instance, their progressive structure may discourage rewards for raising productivity because less and less of the wage increase goes to employees and more to the State. In addition, taxes and social security contributions are often related to the level of income of employees. Hence, they may distort employers' choice of a working time system or deter them from hiring additional staff. For instance, when there is an upper ceiling on the income on which taxes are payable, employers will be deterred from sharing high-paying jobs between two part-time workers, since their combined salary will no longer exceed the ceiling. Such ceilings also encourage the use of overtime rather than hiring additional staff. Taxes in the form of a fixed rate per employee will have the same detrimental effect upon recruitment. Conversely, tax-free floors provide an incentive to hire part-time workers for very few hours per week, thus raising the problem of "involuntary" part-time employment, or rather, of weekly hours of work falling short of employees' preferences.⁽²¹⁾

In the machinery industry, supplements to wages (as a percentage of wages) are highest in Sweden in both electrical and non-electrical machinery (53.0 per cent and 65.4 per cent, respectively, in 1992-94) and lowest in New Zealand (3.3 per cent and 2.3 per cent respectively) (see tables H.1-H.4, The industry in numbers). In most countries, supplements to wages are higher in non-electrical than in electrical machinery, a pattern which became more widespread between 1980 and 1994, except in Denmark and Portugal, where the reverse tendency could be observed. However, one must bear in mind that these supplements include social contributions paid by employers on behalf of their staff plus other allowances, depending on national definitions and statistics. Moreover, statutory social contributions paid by employers in Europe are generally higher than in other developed countries; Japan and the United States rely more extensively upon voluntary contributions. But the figures for Japan understate the social element because part of it is consolidated into wages; for instance, this is the case of family allowances, which in Europe are funded through employers' contributions.⁽²²⁾

Finally, an inverse relation between changes in wages and in supplements could be observed in about half the cases, mainly in non-electrical machinery, which generally has higher labour costs than electrical machinery. This suggests that changes in supplements are, to some extent, passed back to employees. This inversion occurred in the following countries: in electrical machinery, Denmark, Mexico, New Zealand, Norway, Spain and the United Kingdom; and in non-electrical machinery, the same set of countries plus Luxembourg and the Netherlands. This finding is consistent with the conclusions of an EU study of non-labour costs in general, which found that, to a major extent, social contributions paid by employers appear to be passed back to employees: wages are lower than they would otherwise be. Countries with similar levels of labour productivity tend to have similar labour costs, irrespective of the rate of social contribution imposed on gross earnings. Therefore, the respective shares of labour costs in value added also tend to be similar. To the extent that social contributions are in fact borne by employees in the form of lower wages, their position is not much different from that in countries which rely more extensively upon voluntary contributions; moreover, the competitive position of firms is not adversely affected.⁽²³⁾

7.6. Evidence from case-studies

A study of full-time workers in the semiconductor industry in the United States concluded that wage inequalities in the sector have risen.⁽²⁴⁾ This was primarily driven by the following factors: a slowly increasing premium for tertiary education, and a rapidly growing premium for managerial and professional staff -- both of these differentials began to widen considerably in the mid-1980s -- as well as the wage dispersion within each occupation, which increased in the 1990s. But, contrary to national trends, the return to experience has not increased for workers in the semiconductor industry. This might be due to suppression of seniority-related wage increments under new pay systems, to a "new skills" premium in an industry which is marked by rapid technological change, and to the fact that the data do not include non-wage compensations which are common practice in the semiconductor industry, especially for senior employees.

In the United States economy as a whole wage inequalities have increased rapidly in the last 20 years, while mean wages have grown slowly (wage inequalities are quite pronounced when average wages are much higher than the mean, or median wage). In the industry, real wages of production workers increased until the mid-1980s and then decreased thereafter, while those of managers and professionals rose steadily from 1980 to 1992. Much of the increment in real wages of production workers in the early 1980s is probably due to a change in the composition of the labour force. Many of the lowest skilled and most of the labour-intensive jobs were eliminated through a combination of automation and outsourcing overseas. The production jobs that remained were thus predominantly higher skilled jobs and the relative share of production workers in total employment of the industry declined from 60 to 40 per cent over the 1980-92 period.

Wage dispersion in the industry closely tracks that for the economy as a whole. Overall, the difference between the highest and the lowest deciles increased from 260 per cent in 1979 to 412 per cent in 1995; for managers and professionals it increased from 250 to 413 per cent, more than for production workers, for whom it grew from 150 to 267 per cent over the period. However, data from 16 chip-making firms exhibit a wide variation in pay structures. The percentage difference between the top and bottom wages varies from 46 to 258 per cent for operators, 21 to 177 per cent for technicians, and 123 to 241 per cent (possibly 851 per cent) for engineers. As might be expected, engineers have the steepest career path, followed by technicians and operators, although technicians tend to have fewer grades than operators. Furthermore, the ratio of operators' to engineers' starting wages ranges from 0.32 to 0.96, and the ratio for technicians versus engineers ranges from 0.42 to 1.19. Some firms offer their workers long career paths and potential wage growth while others do not. Overall, wage inequality in the semiconductor industry is high and similar to that in the economy as a whole.

An Australian case-study also finds increasing earnings dispersion in the machinery industry as a result of the emergence of enterprise bargaining as the primary locus for wage determination.⁽²⁵⁾ Many of the changes in the Australian wage system at large are the result of developments in the metal and engineering sectors. Since the early 1980s, the wage bargaining process has shifted to "managed decentralism" from a highly centralized system based on awards that are handed down by industrial relations commissions, which set wage floors by means of these awards. Later phases involved reforming classification structures, skill formation arrangements and work reorganization, which proceeded in all other industries along the lines pioneered in the metal industry. Decentralization has divided the workforce into three groups: those remaining dependent on awards, those negotiating collective agreements and those relying on individual contracts. In metal manufacturing, about half of employees are covered by some kind of enterprise agreement, while the other half is presumably mostly dependent on safety net adjustments (SNAs) made under the award system. Those dependent on SNAs benefited from a wage increase in the order of 8 per cent since 1991, while the cumulative impact of wage bargaining rounds has resulted in a wage increment of 26.5-30.5 per cent over the period. Workers at the bottom of the pay scale could at best maintain their position relative to award rates, implying that they have suffered a significant decrease in both real and relative wages because award rates themselves have declined substantially in real terms.

Average pay in Australian metal manufacturing in 1993 was higher than that in manufacturing in general or the economy at large. This was true for all the major blue-collar occupational groups. Within this sector, however, significant differences could be observed. The highest proportion of low-paid employment was found in machinery, electronics and electrical industries, compared to other subsectors in the metal industry, such as transport equipment, metal fabrication and basic metal. Its incidence was also higher than the average for the economy as a whole. Within this sector average wages were higher in electronic and light electrical parts than in heavy engineering plant and equipment, mirroring our earlier finding across countries. But these differences are primarily determined by differences in the occupational profile of these two subsectors; controlling for occupational differences, a small wage advantage remains for heavy engineering plant and equipment. In other words, wages of all occupations are pulled upwards in the high-wage subsector owing to the predominance of high-wage occupations, while the reverse occurs in the other subsector. Gender wage differentials appear to be larger in higher wage occupations and in the higher wage industry, again reflecting OECD patterns. Wage differentials have increased since the mid-1980s, the most marked differences occurring among women employees. Decentralization appears to have increased inequalities between, as well as within, industries and occupations.

A comparative study of career paths and wage structure in the United States and Japan in the electronics, automobile and telecommunication services industries sheds interesting light on the relationship between pay incentives, wage inequalities and productivity improvements.⁽²⁶⁾ It challenges the view that greater inequalities are necessary for faster economic growth. From 1970 to the present, productivity increased much faster in Japan, together with a more rapid growth of workers' real wages. From 1985 to 1991, real wages in Japan rose by 11.2 per cent and productivity by 35 per cent, compared to 1.8 per cent and 4.8 per cent, respectively, in the United States. Yet, overall, there are smaller wage inequalities in Japan, except for the gender gap, but steeper age-earnings profiles and longer career ladders; they provide an incentive to ensure that workers' skills increase with age. In the lifetime employment system management has an interest in ensuring continuous upgrading of their workforce and increases in productivity. These are matched by seniority (or rather life-cycle) pay increments, through a set of institutions that is more structured than in the United States, but which produces smaller inequalities. Thus, career ladders are generally better developed in Japan, but with some variations across industries; age-earnings profiles in large companies are steeper in the Japanese automobile and electrical industries than in the same industries in the United States.

Seniority pay is the main component in the Japanese pay system, and the main means of justifying wage differentials for the same job, followed by job grade and performance. Skills and performance are appraised frequently and tied to promotion and pay -- appraisals are equally important for blue and white-collar workers -- whereas in the United States job appraisals are only loosely tied to pay. The pay rates of team members are more compressed in the United States, which results in free-rider problems. In the Japanese pay system, the individual performance appraisal includes an evaluation of workers' contributions to the team and co-workers; individual and team components are thus made compatible and mutually reinforcing. The appraisal system is also seen as more objective and free-riding is further discouraged by long-term employment relationships. Fast-tracking for highly rated workers was more prevalent in Japan than in the United States, but none of the Japanese companies surveyed had a post-and-bid system (whereby jobs are publicly posted and open for applications); job assignment by management was the norm. Furthermore, in Japan, there is no American-style division between production and craft workers with respect to pay, work and status among blue-collar workers, both groups being integrated into a single pay table; career ladders for production workers incorporate skilled and semi-skilled jobs. High-school and university graduates are on the same pay schedules, but the latter progress much more rapidly. The difference between these two groups is therefore less pronounced than in the United States. Similarly, the differential is also smaller between blue and white-collar workers, as well as between the average production worker and CEOs.

In contrast, in the United States, much of the growth in inequalities in the 1980s occurred within age-education-gender groups. Production and clerical workers usually have highly compressed pay systems; seniority has more importance with respect to job assignments, scheduling and lay-offs than with regard to pay. Age-earnings profiles for blue-collar and clerical workers are relatively flat, although pay differentials tend to be somewhat larger in non-union companies. College-educated workers are in professional and managerial positions with different pay systems and steeper age-earnings profiles than production and craft workers; interfirm mobility, whether actual or potential, provides a major incentive for rapid promotion of exceptional performers.

In addition, the effect of firm size on pay is more important in Japan, especially for women. Gender differences in promotion and pay have decreased in the United States, while the gender wage gap has increased in Japan since the mid-1970s. Finally, the wage structure and increases are bargained nationally for all unions during Shunto, the annual spring wage bargaining round, and the earnings profiles of union workers across industries are therefore fairly similar. Conversely, in the United States, the pay ratio of salaried manager to assembler varies substantially across industries: it was 2.3 to 1 in motor vehicles, 3.1 in electrical machinery, 2.8 in manufacturing, and 2.3 in all industries in 1979. Thus, the study concluded that inter-industry wage differentials seem to be socially rather than technologically determined. This conclusion also points to the cultural dimension of incentive systems and hence to the difficulties involved when attempting to transfer them to another context.

This study of high-performance American firms that emulate Japanese-style management methods further concluded that their pay systems appeared to be still in transition. Innovative pay systems were only beginning to be introduced in some of these firms and on an incomplete basis. There seemed to be a low but increasing incidence of skill-based and performance-related pay as well as of profit-sharing schemes. Pay-for-knowledge was more likely to be found in establishments that have work teams, job rotation, total quality management systems or problem-solving quality circles. Finally, profit-sharing pay appeared to account on average for only about 5 per cent of pay for non-managerial employees in companies with such plans.⁽²⁷⁾ But this proportion is nonetheless perceptible to employees and bears on motivation and morale, as the example of Brimfield Precision shows (see box 7.1). More generally, this comparative study is revealing of the importance of coherently integrating pay systems in the other elements of the employment system.
 

1.  European Commission: Green Paper: Partnership for a new organization of work (Brussels, 1997), p. 9.

2.  "Engineering change", in IRS Employment Trends, 628 (Mar. 1997), p. 10.

3.  "Work organisation and employment in electronics", in European Industrial Relations Review, 257 (June 1995), p. 22.

4.  "Engineering change", op. cit., p. 10.

5.  Association of German Machinery and Equipment Manufacturers (VDMA): Neue Formen der Arbeitsorganisation -- Aktuelle Beispiele aus dem Maschinenbau, Band 2, Lohnmodelle für Gruppenarbeit, VDMA Daten -- Fakten -- Argumente 11 (Frankfurt, VDMA, 1996).

6.  "Engineering change", op. cit., p. 11.

7.  G. Ferrante (FIOM-FIM, Italy): Untitled paper presented at the IMF-ILO Conference on the Machinery Industry, Geneva, 26-28 Nov. 1997, p. 7.

8.  A. Gorgeu and R. Mathieu: Recrutement et production au plus juste: les nouvelles usines d'équipement automobile en France, Centre d'études de l'emploi, Dossier 7 (Gap, France, Louis-Jean, 1995), pp. 91-110.

9.  C. Clausen and B. Lorentzen: "Workplace implications of FMS and CIM in Denmark and Sweden", in New Technology, Work and Employment, 8(1), Mar. 1993, pp. 21-30.

10.  OECD: Employment Outlook, 1997 (Paris, 1997), pp. 1-25.

11.  OECD: Technology, Productivity and Jobs, Vol. 2, Analytical Report (Paris, 1996), pp. 79-102.

12.  OECD: Employment Outlook, 1997, op. cit., pp. 1-25.

13.  OECD: Employment Outlook, 1996 (Paris, 1996), pp. 59-108.

14.  Wages here correspond to earnings according to ILO concepts and definitions; they include the base wage, overtime payments and other bonuses but not employers' contributions to social security and pension schemes. Labour productivity is calculated by the OECD the same way as was done in Chapter 2 with UNIDO data, namely as real value added per employee, rather than per hour, because consistent data on labour hours are not available.

15.  Labour's shares reported by the OECD are calculated on the basis of labour compensation, which is a more comprehensive measure of labour costs than earnings, and they are therefore larger than the labour's shares discussed in Chapter 2; the latter have been calculated with UNIDO data on the basis of earnings.

16.  Labour's shares are similar, but not identical, to unit labour costs. The latter are calculated as the ratio of labour compensation in current prices to value added in constant prices (i.e. adjusted for inflation). Therefore, when there is inflation, unit labour costs will increase even though labour's shares might remain the same. In table H, The industry in numbers, this is reflected in an increase in unit labour costs over the period 1980-94 even when corresponding labour's shares were declining.

17.  The industrial classification system (SIC) of the Bureau of Labor Statistics differs somewhat from ISIC, which was used throughout this report.

18.  European Commission, DG V: Employment in Europe, 1994 (Luxembourg, Office for Official Publications of the European Communities, 1994), pp. 140-142.

19.  K. Zühlke-Robinet and S. Lehndorff: Company policies with regard to the reduction of working time: Germany, Working Paper No. WP/97/51/EN (Dublin, European Foundation for the Improvement of Living and Working Conditions, 1997), pp. 39-40.

20.  "Permanently temporary: High-tech firms rely on new working class", in International Herald Tribune, 31 Mar. 1998, pp. 1 and 15.

21.  European Commission, Green Paper, op. cit., p. 11.

22.  European Commission, DG V: Employment in Europe, 1994 (Luxembourg, Office for Official Publications of the European Communities), 1994, pp. 129-132.

23.  ibid., pp. 129-140.

24.  B.A. Campbell and V.M. Valvano: "Wage structure in the electronics industry", in C. Brown (ed.): The competitive semiconductor manufacturing human resources project: Final report (Phase I), Report CSM-38 (Institute of Industrial Relations, University of California at Berkeley, Dec. 1997).

25.  J. Buchanan: The impact of flexible labour arrangements in the Australian metal and engineering sector with special reference to developments in the machinery, electrical and electronic industries, Sectoral Activities Programme Working Paper (Geneva, ILO), forthcoming (Chapter 7).

26.  C. Brown, Y. Nakata, M. Reich and L. Ulman: Work and pay in the United States and Japan (New York and Oxford, Oxford University Press, 1997), pp. 97-136.

27.  Survey carried out by Towers, Perrin reported in S. Pearlstein: "Employers reluctant to share profits", in San Francisco Chronicle (5 Sep. 1994), p. D1 and cited in C. Brown et al., op. cit., p. 103.

SECTORS | ACTIVITIES | MEETINGS | PUBLICATIONS