Jump to ContentJump to Main Navigation
The Japanese Employment SystemAdapting to a New Economic Environment$

Marcus Rebick

Print publication date: 2005

Print ISBN-13: 9780199247240

Published to Oxford Scholarship Online: July 2005

DOI: 10.1093/0199247242.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 24 November 2020

Industrial Relations

Industrial Relations

(p.75) 5 Industrial Relations
The Japanese Employment System

Marcus Rebick (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Examines recent developments in industrial relations in Japan. The decline in union density is analysed and found to be due to a mixture of causes including sectoral shifts in employment and the failure of unions to organise new workplaces. The weakened role of the Shunto in the early 2000s and the increase in civil litigation and court-mediated dispute resolution are symptoms of the declining role of unions. Japanese labour relations are becoming more individualised.

Keywords:   dispute resolution, employment, industrial relation, Japan, litigation, union density, unions


Like other economic institutions, the industrial relations system in Japan has seen significant developments throughout the period since the end of the Second World War. The main features of the system were well established by the end of the 1950s, but the slowdown in growth and the need to restrain inflation in the mid-1970s led to a major change in the behaviour of the union movement. The 1980s and 1990s saw a consistent decline in union density and the deflationary period of the past few years has seen a major change of strategy of the main union federation, Rengō. At the beginning of this chapter I will sketch out some of the historical background before moving on to discuss the changes and challenges facing the industrial relations system.


Japanese industrial relations did not, of course, begin after the Second World War, but the war does provide an important break with the past as liberalization and democratization took place under the authority of the American Occupation. Prior to the war, the nascent union movement had been repressed to a great extent, through legislation such as the Public Peace Law of 1900, which was used by authorities to break up strikes and other gatherings of unions. Union membership never rose above 7% of the workforce during this period. In 1938 the government outlawed unions and replaced them with factory councils known as Sanpō (Industrial Patriotic Associations), with representatives from management and labour. Although these councils were not always effective in managing labour disputes, they did leave a legacy for the future. Management and government did, at least, pay lip-service to the concerns of workers, and their need for dignity (Gordon 1985). The councils also forced management to sit down with labour representatives, and this provided a precedent for future cooperation. The Sanpō were disbanded in 1945, but a new union movement immediately sprang up in their place.

The Occupation Reforms

The American Occupation introduced a number of democratizing measures in the early aftermath of the war. One of these was the introduction of legislation (p.76) to promote the development of an industrial relations system. The right to organize is explicitly laid out in the Japanese constitution as a basic human right that is ‘eternal and inviolable’. The first Trade Union Law was introduced in 1945. The response to this legislation was very rapid and by the end of 1946, 46% of employees were organized, rising to a peak of 56% in 1949 (Shirai 1983b: 140).

The Absence of an Exclusivity Clause

To a great extent the Trade Union Law is modelled on American labour law, especially the National Labor Relations Act and Taft-Hartley Act, but there are a number of differences. Here I would like to draw attention to one in particular. American law provides for the designation of a bargaining unit for which workers can elect an exclusive representative by an election with a simple majority. In Japan, there is no such designation of a bargaining unit and there may be more than one unit representing any given workplace or other group of workers. This is one example of the way in which the Japanese have modified the original intentions of the Occupation Authorities. Neither employers nor the unions were in favour of an exclusivity clause in the labour legislation. The employers were against this because it obviously put the union representatives in a more powerful position. The unions tended to be against it because the union movement was divided and different groups feared that exclusivity might lock them out of the workplace. In addition, it would be difficult to establish the appropriate bargaining units as there were not the same demarcations of jobs in Japan as in other countries.

The result has been that the union movement has been left more divided than it otherwise might have been, and the unions themselves have been in a much weaker position within firms. In the event that a militant union organizes some of the workers, it is possible that a second (or third) union taking a more moderate stance will also be able to organize employees, often with the tacit approval of management. The famous strike at Nissan in 1953 is the textbook example of such a manoeuvre on the part of the employers, anxious to put down an antagonistic union (Cusumano 1985). The lack of exclusivity also made it more difficult for unions to be organized on an industry-wide basis and led to the development of more cooperative unions that were mainly interested in promoting the success of the enterprise.

The Roll-Back

The American Occupation also acted to roll back some of the provisions of the original legislation in the late 1940s. The most important change was the introduction of the Public Corporation and National Enterprise Labor Relations Law of 1948 that denied public-sector workers the right to strike. This remained a contentious piece of legislation until the 1980s, at which point most of the public corporations (railways and telecommunications) were privatized. The onset of the Cold War led to a push by the Occupation authorities to purge communists from the government. The end of the 1940s saw the introduction of tight monetary policy under Joseph Dodge, and the resulting recession (p.77) gave management the opportunity to fire many of the most radical workers.1 By the late 1950s much of the postwar industrial relations system had been established. Management and labour had struck a bargain whereby employees were given job security. In return, management had the right to allocate workers where it chose in the workplace, and the kind of job demarcations that are characteristic of other industrial relations systems did not develop. Needless to say, part-time workers were not included in this bargain.

The Labor Relations Adjustment Law

Along with the trade union laws, the Labor Relations Adjustment Law of 1946 (amended in 1949 and 1952) is of central importance to labour relations in Japan. Although grievances can ultimately be taken to the courts, the first route to settlement is through Labor Relations Commissions, established in every prefecture and at the national level under the Trade Union Law. These bodies act as conciliators or mediators in disputes, helping parties to resolve differences without recourse to litigation.

The Shuntō

The development of the postwar system reached completion with the introduction in 1954 of the ‘Spring Bargaining Offensive’ or shuntō. Starting with a core of left-wing unions in the Sōhyō federation, the movement quickly blossomed, incorporating some 4 million workers by 1961 (Koshiro 1983). In the shuntō process, unions in leading industrial sectors set the pace early in the year, asking for a common wage increase, to which employers respond with a counter-offer. Other industries then follow suit and although there may be some deviation from the common proposed wage increase, it serves as a useful benchmark. During the first 20 years, different industry groups were responsible for taking the lead in making settlements, but since 1975, the unions in the metalworking industries (IMF-JC) have set the pace. The IMF-JC unions have tended to be the most moderate and cooperative in their approach, as their industries must compete in the international market. In general, their approach is to push for wage increases that are based on productivity gains, and critics maintain that they have not pushed hard enough.

The union movement has been divided throughout the postwar period. At the most basic level, collective bargaining takes place at the enterprise level, although pattern bargaining does lead to coordinated settlements. Enterprise unions are in turn organized into industry federations which may, in turn, be subsumed under a national federation. There have been many different national federations over the past 50 years and Figure 5.1 traces their development and genealogy. The complexity of the diagram reflects the degree of disunity and weakness of the various federations. Since 1989 there has been greater stability. Today, the main federation is Rengō with over 7 million members out of a total of 11.5 million union members. Two other, more left-leaning federations are Zenrōren with roughly 1 million members and Zenrōkyō with roughly a quarter (p.78)

                      Industrial Relations

Figure 5.1. The development of postwar labour organizations in Japan

Source: Adapted from JIL (1992: 50–1).

(p.79) of a million members. Rengō was formed by the merger of two different federations, Dōmei, which was the more centrist federation, comprising many of the manufacturing industries that have taken the lead in the shuntō, and Sōhyō, a more left-wing federation to which most public-sector unions belonged. In general, public-sector workers have tended to be most left wing in their approach, and since they are not under competitive pressures, they also tend to argue for higher wage settlements. The same can also be said of private sector workers working in industries such as railways that don't face international competition.

The Relations with Political Parties

The relationship of the union federations to the political parties has also had a mixed history. Sōhyō was heavily allied with the Japan Socialist Party, while Dōmei tended to favour the more moderate Democratic Socialist Party. After the founding of Rengō in 1989, Rengō tried to develop its own political party, similar to the labour parties of other countries, but without success. Recent attempts by the Democratic Party, the leading opposition party in Japan, to get support from Rengō have not been successful. Again, one has the impression of weak cohesion within the union movement.


Union density has been in decline in Japan since the oil shock of 1973, falling from 35% to less than 20% of non-agricultural employees by 2003. The decline in enterprise unionism is one of the main ways in which the employment system is changing. There are a number of explanations for the fall in density and here I will briefly point to some of the factors that have been important.

Change in the Composition of the Labour Force and the Structure of Industry

The decline in manufacturing, the increase in the number of women working as employees and the increased use of contract and part-time workers may all have contributed to the fall in density, if we assume that the propensity for any type of worker to be unionized remains constant. Table 5.1 shows that the drop in unionization rates between 1975 and 2002 has taken place within every industry except for construction. If we hold the industry unionization rates of 1975 constant and look at the change in the share of each industry in employment we find that only 3.3 percentage points out of the total 13.6 point drop can be explained by shifts in industrial composition. The remaining 10.3% drop must occur within industries, partly as a result of the feminization of the labour force and even more as a result of the greater use of part-time workers (most of whom are women). In order to account for these trends, I conduct a more comprehensive shift-share analysis to derive a counterfactual density that accounts for (p.80)

Table 5.1. Change in unionization rates by industry, 1975–2002 (in per cent)





















Finance, insurance and real estate




Transportation and communications
















Source: MHLW2 (1975, 2003).

(1) changes in industry structure, (2) changes in the proportion of employees that are women, and (3) changes in the proportion of employees that work part time.2 The result is that overall density should have fallen from 34.5% to 27.9% if the propensity of full-time workers to be unionized had remained constant. So consideration of all three factors now accounts for 6.6 points, or nearly half of the 13.6 point drop.

There are some other minor factors that will also have led to a decline in union density. For example, the ageing of the workforce has meant that firms have tended to increase the number of managerial posts per employee in order to maintain morale. Japanese union law prohibits employees of managerial rank from being members of unions, so this may have had a minor impact on total union density. A second factor that has had some influence is the decline in the proportion of employees that work in large firms. These two factors combined should account for less than one percentage point in the drop in density. Around half of the decline in the unionization rate remains unexplained.

The Decline in the Rate of New Union Births

The explanation for the remaining part of the drop in union density lies in the fact that new unions are not being organized. Figure 5.2 shows the numbers of union members in new unions and their share of all employees. It is clear from the chart that organization of new unions fell off dramatically after 1975. In fact, since a decline in union membership has been taking place at the same time (partly due to the dissolution of unions) the rate of unionization has dropped consistently since then.

A number of explanations have been put forward to explain this decline in the organization of new unions. Freeman and Rebick (1989) attribute the decline in (p.81)

                      Industrial Relations

Figure 5.2. Membership of newly organized labour unions and share of employees 1966–2003

Source: New members: MHLW2; employees: PMO2.

part to management resistance to the formation of new unions as profit margins have fallen since 1975. This is likely to be only part of the story, however. Another equally, if not more, important factor is that there is little interest in unionization when it does not deliver much in the way of better pay or working conditions (Tsuru and Rebitzer 1995). Most of the non-unionized firms follow the larger, unionized firms in determining their pay settlements. The result is, that, once firm size is taken into account, there is no discernible union-wage differential (Tachibanaki and Noda 2000).3 Japanese union dues are relatively high at 2–3% of monthly pay and this also discourages potential members from joining.

Alternatives to Unions

The absence of unions does not necessarily mean that workers have no voice in the workplace. One such institution is the joint labour-management committee (JLMC) which exists in both unionized and non-unionized firms. Kato (2003) reports that some 80% of publicly traded firms had JLMCs in 1993. Among smaller firms these institutions are found in 90% of unionized firms, but in fewer than half of non-unionized firms (Sato 1997). The JLMC is primarily used for information-sharing and as a means by which management can consult with employees on a range of issues, including layoffs, fringe benefits, working hours, and pay. Management will often ‘open its books’ to the employee representatives, (p.82) sharing information on sales, production, and even on plans for new product development. In unionized firms, these practices can smooth the bargaining that takes place before the annual shuntō (Morishima 1991b) and can raise labour productivity (Morishima 1991a). Kato (2003) notes, however, that there has recently been much less consultation in non-unionized firms on transfers, layoffs, and pensions and he suggests that unions may prevent JLMCs from becoming dormant, or else simply venues for management to announce personnel policy. If this is the case, then the decline in unionization rates will also weaken the role of JLMCs across the economy.

In addition to JLMCs, or as an alternative, many firms have non-union employee organizations. Surveys of employee organizations in small and medium-sized firms show that although they are often formed to organize leisure activities, roughly one in five non-unionized firms appears to have an organization that will discuss substantive workplace issues with employers (Nitta 1992; Sato 1997). In the majority of these cases, the main issues discussed are working hours, holidays, and safety and health.

One final alternative voice mechanism is to approach one's supervisor, possibly over after-hours drinks. Morishima (1999) notes that one of the problems facing the union movement is that this is the preferred method of dealing with grievances, even in unionized organizations. He cites one survey where 60% of employees indicated that they preferred to talk with their supervisors to air grievances. Only 15% would approach their labour union or employee organization.


The 1990s have brought additional challenges to the union movement as the long recession has made it difficult for unions to gain much in wage settlements during the shuntō. Table 5.2 shows that the average shuntō settlement for large firms has not declined in real terms over the past 10 years. The average wage settlement in the 1990s showed a 2.5% increase: not much lower than the average of 3.1 in the 1980s. Bonuses tell a different story, however. There has been an average decline in real terms in bonus payments over the past decade. In the case of both the wage increase and bonus increases that are reported, the figures show the average wage increase (increase in the wage bill divided by the number of employees) for the firm. For firms that have an ageing employee structure, this wage increase will largely be eaten up by the increases in wages that tend to accompany age and seniority. Thus, the actual increase from the point of view of the employee's lifetime earnings tends on average to be about 2% lower.4 The average real increase in the 1990s for large unionized firms was around 0.5% per annum, and bonus payments fell by more than 2% per annum in real terms, once age- and seniority-based rises are taken into account.

The union membership has, however, been far more concerned about job security than pay, and willing to make major concessions (especially in bonuses). Again, unions in the export industries have been at odds with the public-sector (p.83)

Table 5.2. Annual wage and bonus increases for large firms, 1991–2003


Nominal wage increase

Real wage increase

Standardized interquartile range1

Rate of increase of bonus (real2)


































































Notes: (1) The standardized interquartile range is the difference between the third and first quartiles, divided by twice the median.

(2) Average of summer and winter bonus increases.

Source: Wage increases from JPC (2004), table C-33. Values for approximately 290 of the largest, listed, unionized firms. Bonus figures from MHLW4 (2004), table E-8, large firms only.

unions. Towards the end of the 1990s, the solidarity of the shuntō was broken both by individual firms and by industry federations. For example, Weathers (2001b) points out that the Federation of Steel Workers had no wage increase in 1995 and only token increases in the next two years. The electronic workers’ federation Denki Rengō also began to delink its wage settlements from other sectors as firms began to individualize pay settlements. Finally, for the Spring Offensive of 2003, Rengō announced that it would no longer seek a unified wage increase, but would allow individual industry federations the freedom to negotiate separate wage settlements. Instead, the federation has now embarked upon a new strategy of ‘minimum standards’.

Minimum Standards and the 2003 Spring Bargaining

The minimum standards advocated by Rengō are mainly based on the first decile of the wage distribution in 2002. Using the age- and seniority-earnings curves, standards are set for age levels up to age 35 and tenure up to 17 years. In addition, targets are set based on the average levels of 2002. Some examples of the minimums and targets are shown in Table 5.3.

The minimum for 'standard employees’ with the maximum number of years of seniority is determined by looking at the first decile of wages reported by member unions. The minimum for newly hired employees is based on the average of the minimum earnings in existing collective agreements. The targets for ‘standard’ (p.84)

Table 5.3. Rengō's minimum standards and targets—Spring Bargaining of 2003
























Source: Rengō Internet web page.

employees are based on the average of wage settlements for skilled production workers in the unions surveyed.

The most important feature of this scheme is the desire on the part of the unions to maintain the wage curve up to age 35, including the seniority component. This seems to be the most important concern of employees. At the time of writing it remains to be seen whether or not this strategy will be successful. In addition to specifying minimums for standard employees, minimums and targets are also given for part-time workers. At 790 and 900 yen per hour respectively, they are substantially higher than the minimum wage mandated by the government, which is highest in Tokyo at 708 yen per hour. Part-time workers are barely represented in Japan (only 3% were organized in 2002), but the increasing use of part-time workers by firms means that standard employees must also take account of their wages.

The unions are also pressing for limitations to the use of unpaid overtime work and for increased use of vacation days by employees. Although working hours fell in the early part of the 1990s recession, unpaid overtime has increased as firms attempt to cut labour costs. It is very difficult for firms to force employees to go home if they feel that their career (or even their job security) with the firm depends on putting in extra work. In part, this problem is also related to the nature of work organization. Teamwork may mean that employees may need to wait for others to finish before their own work can be completed. This also makes scheduling vacations more difficult as any individual's absence may have a major impact on the work of others.


One way to gauge the overall effectiveness of the labour movement is to look at trends in labour's share of national income. Figure 5.3 shows trends in the proportion of distributed national income that is in the form of employee compensation. Since the self-employed sector is in decline, we would expect this rate to be rising, other things equal. Therefore, I also show an ‘adjusted labour share’ (p.85)

                      Industrial Relations

Figure 5.3. Trends in labour's share of national income, 1970–2002

Notes: Labour's share is the proportion of distributed national income that is paid to employees. This is divided by the proportion of employees among those employed to get an adjusted estimate of labour's share. This attributes the same average level of labour income to non-employees as to employees.

Source: JPC (2004), table A-1.

Table 5.4. Trends in the number of labour disputes


Number of disputes

Number of participants

Percentage of disputes where no industrial action was taken





























Source: MHLW4 (2004), table I-12.

which divides the ratio by the proportion of employed who are employees, normalized at unity in 1990. Using the adjusted series as a guide, we see that labour's share rose rapidly in the early 1970s, a period of intense activity on the part of the labour movement (Table 5.4). Since 1975, however, and the advent of the more cooperative Dōmei unions, labour's share fell until 1990 before levelling out. It has risen slightly since 1990, more as a reflection of poor profit levels than of any real gains on the part of labour.


Industrial relations have been quiet in the past few decades and the number of disputes has fallen sharply in the 1990s, reflecting the weak state of the labour market. The weakness of labour's position is reflected not only in the falling number of disputes, but also in the fact that a rising percentage of disputes are settled with no action taken on labour's part. The last few years of the 1990s also saw a fall in the proportion of disputes concerned with wage levels and a relative increase in those concerned with the firing of workers (Table 5.4).

The fact that unions have been quiet does not reflect the true state of employee–employer conflict, however, as there has been a steady increase in the amount of litigation that has taken place, mostly reflecting individual law suits over dismissal. The increase in such litigation may be seen as a reflection of the declining role of the Central Labor Commission and of labour unions in dispute resolution. Most of the increase in civil litigation has occurred within the category of the ‘regular procedure’, which involves a fuller treatment and is regarded as more serious than the preliminary injunction process (Nakakubo 1996). Most of the cases are brought by individuals, and most relate to wages, but dismissals also feature prominently.

Civil litigation is a cumbersome way for disputes to be resolved and the average time needed for a court decision in the first instance is more than one year. If the case is taken all the way to the Supreme Court, the entire process can take decades. The increase in the numbers of new cases began with the downturn in the economy in 1992, but, as indicated in Table 5.5, the growth has continued

Table 5.5. Civil suits initiated in district courts under the regular procedure, 1990–2002


Number of suits



























Source: Hōsō Jihō (1995, 2003).

(p.87) throughout the last decade. Given the difficulties involved in bringing a suit to the courts, this increase is probably a reflection of much greater levels of underlying dissatisfaction. It may also reflect the ongoing individualization of labour contracts and the decline of the importance of unions in general.

Although the introduction of individualized performance-related pay is a cause of considerable anxiety for employees, labour unions have not been very successful in influencing managerial decisions concerning the implementation of these plans. Morishima (1999) suggests that unions need to take the initiative by actually promoting the introduction of pay-for-performance schemes, but in a manner that allows for the greatest transparency in the evaluation process and that prevents abuse by managers. He admits, however, that this will not be easy, as the increasingly white-collar workplace has a greater diversity of employee interests than in the past, in part due to greater specialization of employees.


Enterprise unionism was held to be one of the central pillars of the Japanese industrial relations and employment system. Although it is still a presence and the shuntō hobbles on in modified form, the rise of individual grievance cases in the courts suggests that, as in other industrialized countries, the ability of unions to successfully fight downsizing in the current economic climate is limited. Furthermore, unions have had little success in organizing part-time workers. The increased use of part-time workers undermines the bargaining power of the union movement in addition to its effect on union density. Finally, on the positive side, it should be noted that the union movement continues to provide considerable wage flexibility. Although there are some critics who believe that high wages may bear some responsibility for the recent rise in unemployment, labour's share declined in the 1980s to levels not seen since the early 1970s. There are other, more convincing explanations for the rise in unemployment. It is to this subject that I now turn.


(1.) It was this kind of action that led directly to the courts providing better job protection. The development of the four required conditions for dismissal (see Chapter 2) was intended to prevent firms from undertaking this kind of purge in the future.

(2.) I compute how much the total density would have dropped if we assume that part-time workers have a unionization rate of zero and then compute the male and female union densities for full-time workers separately for each one-digit industry for 1975. As Labour Force Survey data is used, full-time is defined here as working more than 35 hours per week. I then apply these 1975 densities for full-time workers to 2002 figures for the numbers of full-time men and women in each industry and compute the ‘counterfactual’ number of union members that would exist if the industry-specific full-time worker densities had remained constant. This figure is then divided by the total number of employees to derive a counterfactual density that accounts for (1) changes (p.88) in industry structure, (2) changes in the proportion of employees that are women, and (3) changes in the proportion of employees that work part time.

(3.) There is evidence that unions do have an effect on benefits and working conditions (Tachibanaki and Noda 2000; Nakamura et al. 1988), but this may not be generally well understood.

(4.) JPC (2001, table C-38) makes it clear that expected rises contributed about 2% on average to wage increases in the 1990s.