Union Politics in Hard Times:
The Restructuring of "Japanese-Style Employment Relations" and the Internal Politics of the Japanese Labor Movement in the 1990s

Internal politics within labor movements in bothindustrialized and developing countries have become activein the last decade. In the former countries, labor hasexperienced corporate restructuring and the spread of"lean production," and internal debates within labormovements have become active over how unions shouldrespond to the increasingly hostile environments (Moody,1997: 197-200). In the latter countries, unions havebecome more independent from the control of the state andpolitical parties as a result of democratization ofpolitical systems, and the newly-formed independent unionshave begun to challenge the state- or party-controlledunions (Ibid.: 208-224). Japan is where "lean production"originated: and yet, not much has been written on internalpolitics within the Japanese labor movement.

In this paper, I examine internal debates within the unionmovement in Japan since the early 1990s in response to theincreasingly competitive economic environments. The papersheds light on political dynamics within the labormovement by examining how cleavages of interests have developed between labor unions in different sectors as well as within labor unions. I am going to consider whether the internal politics of the labor movement have become active in response to changes in its contexts, as in the case of labor movements in many other countries.

The analytical framework of the paper is informed by recent studies of the rise and decline of neo-corporatist arrangements in Sweden and other northwestern European countries. Series of articles by Peter Swenson and Jonas Pontusson on this topic offer two important perspectives in analyzing labor politics: (1) conflicts of interests are not necessarily class-based and may develop within labor and capital, and labor and capital in one sector may form "cross-class alliances" against those in another sector; (2) capital plays a crucial role in the rise and decline of institutional arrangements of industrial relations in the sense that the stability of institutional arrangements depends on the support of capital for such arrangements (the withdrawal of capital's support considerably undermines the stability of institutions) (Swenson, 1991, 1992; Pontusson, 1995; Pontusson and Swenson, 1996). On the basis of these perspectives, I am going to examine how sectoral interests have affected the internal politics within the labor movement in Japan as well as how employers' policies have affected institutional arrangements of industrial relations. In addition to the two perspectives, I include the third perspective in the analytical framework, namely that tensions may develop not only between unions, but also within unions over unions' responses to restructuring efforts of employers. As Kim Moody and other labor scholars point out, "rank and file movements" led by union activists and lower-ranking union officials have developed within industry unions in the U.S. and other industrialized countries (Moody, 1997: 197-200; Turner, 1992; Yates, 1993). Tensions seem likely to develop within unions when union leaders, as a part of "cross-class alliances" with employers, sacrifice short-term interests of union members in favor of long-term interests of union organization. My analysis of the Japanese case therefore includes the question of whether there are "rank and file movements" within those unions which form "cross-class alliances" with employers.

In the following, I will first examine the development of the labor movement from the mid-1970s to the late 1980s, focusing on the "unification process" of rival national confederations leading to the formation of Rengo (the Japanese Trade Union Confederation) in 1989. Second, I will discuss the context of the labor movement since the early 1990s, focusing on two issues: (1) a drastic reform of the employment relations to cope with the severe recession, and (2) the deregulation of labor relations through the revision of the Labor Standards Law. Third, I will analyze the internal politics of the labor movement in response to the restructuring efforts of employers and the government. I will show the development of two types of cleavages within Rengo. One is between unions in large, competitive firms on the one hand and unions in small, less competitive firms on the other. The other type of cleavage is between private- and public-sector unions. I will also briefly discuss a "rank and file movement" in the electronics industry that challenge union leaders who tend to take cooperative policies toward restructuring efforts of employers. Finally, I will conclude this paper with some observations on possiblechanges of the labor movement in the future.

Until the late 1970s, the labor movement in Japan hadbeen divided ideologically and politically between twomajor confederations, Sohyo (General Council of TradeUnions of Japan) and Domei (Japanese Confederation ofLabor). Besides these rival confederations, there weretwo smaller confederations, Churitsu Roren (Federation ofIndependent Unions of Japan) and Shin Sanbetsu (NationalFederation of Industrial Organizations). Sohyo was theleft-wing confederation and advocated a class-based labormovement. Domei was the right-wing confederation andadvocated a reformist labor movement and cooperativerelations with management. While Sohyo mainly representedmilitant public-sector unions such as unions in thenational railways and postal services, Domei's membershipwas predominantly moderate private-sector unions. Interms of party affiliations, Sohyo supported the JSP(Japan Socialist Party), and a minority of its memberunions supported the JCP (Japan Communist Party). Domeisupported the DSP (Democratic Socialist Party).

In the late 1980s, the four national confederationswere unified into Rengo in two stages: private-sectorunions were unified into Rengo in 1987, and public-sectorunions joined Rengo two years later. Why were theideologically- and politically-divided Sohyo and Domeiunified into Rengo? The role played by private-sectorunions was crucial for the unification process. Since themid-1970s, enterprise unions of large private-sector firms and their industry-level federations had increased theirinfluence in the labor movement, and these "big businessunions" affiliated with Sohyo, Domei, and otherconfederations started discussions about the formation of a unified confederation. Although leaders of these private-sector unions advocated economic unionism and cooperative relations with employers at the enterprise level(1), they felt it necessary for unions to overcome the limit of enterprise-based activities by extending their activities into the political arena. This was because private-sector unions in export-oriented industries felt the impact of economic recession since the mid-1970s and shifted their strategy from demanding wage increases through enterprise-based wage negotiations to demanding government policies to restrain inflation and to protect employment security. Union leaders aimed at increasing unions' influence over the policy-making of the government and at establishing "Japanese-style neo-corporatism" through the unification of the divided union confederations (Shinoda, 1997).

In 1976, 16 industry-level federations of private-sector unions formed Seisui Kaigi (Trade Union Council for Policy Promotion) as a forum for participating unions to coordinate their activities in making policy demands on the government and as a preparatory body for the possible unification of confederations in the future (Hyodo, 1997: 417). In 1982, Seisui Kaigi was succeeded by Zenmin Rokyo (Japan Private Sector Trade Union Council), a loose confederation consisting of 41 industry-level federations of private-sector unions. Finally in 1987, national confederations except Sohyo were dissolved, and Rengo was formed as a formal confederation consisting of 66 industry-level federations of private-sector unions. Sohyo continued to exist as the confederation of public-sector unions but was dissolved in 1989 as these unions joined Rengo.

The unification movement did not proceed without conflicts among and within unions. Although the two major confederations, Sohyo and Domei, started to express their support for the unification of the labor movement in principle by the late 1970s, they disagreed on how the unification should proceed in practice. One of the important points of disagreement was how to deal with those unions which advocated class-based rather than cooperative economic unionism, particularly those unions which supported the JCP. Domei insisted that the unification process should be led by private-sector unions based on cooperative economic unionism, and that militant unions should be excluded from the unification process. Sohyo, reflecting the interests of leftist unions among its members, insisted on the inclusion of all unions in the unified confederation regardless of their policy orientations. However, since the power balance between the two confederations was in favor of Domei, the unification process proceeded in Domei's terms (see Inoue, 1997). Within Sohyo, tension developed between the leadership and leftist unions. While the leadership was oriented toward making a compromise with Domei, leftist unions opposed the unification process. Such tension, however, did not lead to serious confrontations because pro-JCP unions and pro-JCP members who withdrew from Rengo-affiliated unions formed their own confederation called Zenroren (the National Confederation of Trade Unions) in 1989. In the same year, left-socialists formed another confederation called Zenrokyo (the National Trade Union Council).

The balance of power among the three confederations was such that internal politics within the labor movement, by the end of the 1980s, seemed to be settled in favor of cooperative economic unionism at the enterprise level and the neo-corporatist-style policy participation at the national level. Although Rengo did not organize all unionized workers, its position within the labor movement was dominant. The membership of Rengo in 1990 was about 7.6 million, or 62 percent of unionized workers. The membership of Zenroren in the same year was 835,000, or 6.8 percent of unionized workers. Zenrokyo was even smaller: its membership in 1990 was 290,000, or 2.4 percent of organized workers.(2) The rest of unionized workers (32 percent of organized workers) were not affiliated with any confederation. Table 1 shows the number of organized workers by affiliation from 1988 to 1996.

The dominance of Rengo within the labor movement is reinforced by the support of employers and the government. Nikkeiren (Japan Federation of Employers' Association), the employers' association in charge of labor relations, saw the formation of Rengo as the culmination of the shift in labor relations from the one based on confrontation and power to the one based on dialogue and mutual trust. Thus, the employers' association "welcomed" the unified confederation as a "partner of labor-management dialogue." However, it virtually ignored the formation of the two other confederations (Nikkeiren Times, November 26, 1987 and November 23, 1989). The government also takes more favorable attitudes toward Rengo than to the two other confederations. For example, the government has appointed labor representatives to labor arbitration committees as well as to advisory councils of the government only from Rengo and has not appointed anyone from Zenroren and Zenrokyo. Zenroren criticizes the government for taking the policy of virtual non-recognition toward the confederation (Saeki, 1992: 138-9).

The dominance of Rengo, however, does not mean that the labor movement since the late 1980s became de-politicized under the hegemony of big business unions. The labor movement has faced more hostile economic environments than before since the early 1990s due to the economic recession. Particularly important for the internal politics within the labor movement is the fact that employers have begun to withdraw their support for the existing model of employment relations, on which the ideology of cooperative economic unionism advocated by big business unions has been based. Because of the uncertainty of "Japanese-style employment relations," some unions, particularly those representing workers in small- and medium-sized firms, have begun to question the dominance of big business unions in the labor movement.

When Rengo was established in 1989, its leadership expected that labor would increase its influence in the economic as well as political arenas based on the enhanced organizational unity of the labor movement. As a part of the confederation's strategy for greater influence, it started an organizing campaign in 1990 in order to increase its membership from 7.6 million to 10 million by 1993. However, Rengo's membership not only failed to grow as planned (the membership in 1993 was 7.8 million) , but also started to stagnate in the mid-1990s (see Table 1). Moreover, the labor movement as a whole suffered from the decline in union density: union density fell from 30.8 percent in 1980 to 25.2 percent in 1990, and to 23.2 percent in 1996 (see Table 2).

What explains the difficulty of Rengo in organizing new members and the decline of union density? Tsuyoshi Tsuru and James Rebitzer, based on a survey conducted in 1992, argue that the decline of density is not due to "changes in the sectoral composition of employment," but due to the absence of positive effects of unions on wages and benefits that "is likely to limit severely the attractiveness of unions to the non-union work-force" (1995: 465, 482). The absence of union effects on wages and benefits is in turn explained by the enterprise-based structure of the Japanese labor movement. As they put it:

The identification of the union with the enterprise...limits the willingness of the union to seek wage gains that might hinder the competitiveness position of the company. Should profits come under pressure (as happened in Japan as a result of the oil shocks in the 1970s and the heisei recession in the 1990s), it is natural for enterprise unions to adopt an accommodating position [thus reducing the attractiveness of unions to non-union workers] (ibid.: 482).

Their study suggests that the enterprise-based structure of the labor movement makes unions vulnerable to the logic of economic competition under adverse economic conditions, which in turn reduces the incentive of non-union workers to form or join unions. In other words, Rengo's difficulty in expanding its organizational base reflects the recession of the Japanese economy since the early 1990s.

The economic recession has caused not only the stagnation of union growth, but also changes in employers' attitudes toward "Japanese-style employment relations." The model of "Japanese-style employment relations," consisting of three main components (employment security, seniority wages, and enterprise unionism), has been widely seen as one of the important reasons for the economic success of Japanese manufacturers in global markets in the 1970s and 1980s.(4) Although employers saw seniority wages as a potential cause of economic inefficiency, they tried to solve the problem of potential inefficiency not by abolishing seniority wages, but by gradually increasing the portion of ability-based wages at the expense of seniority wages (Kyotani, 1995: 6). The labor movement, particularly unions in large private-sector firms, had come to support this model of employment relations by the end of the 1960s. In this way, "Japanese-style employment relations" were institutionalized mainly in large private-sector firms and remained a relatively stable feature of industrial relations even during the world-wide recession due to the oil shocks in the 1970s and the recession due to the appreciation of yen in the 1980.(5)

The economic recession in the 1990s, however, is much severer than the previous ones, and employers have found the existing system of employment relations increasingly inflexible in coping with competitive economic environments. In 1992, Nikkeiren started proposing a drastic reform of wage system and the division of regular employees between the "flow" and "stock" types (Nakanishi, 1995: 13). In 1995, the employers' association further elaborated its reform plan in a document titled "Japanese-style management in a new era." The document, adopted by the association's convention as an official guideline for personnel policies of its member firms, proposed the following policies among others: the division of regular employees whose employment status is currently secure into three groups with different duration of employment,(6) the shift from seniority-based wages to wages strictly based on ability and performance, the introduction of discretionary work scheme to wider categories of white collar work, and the maintenance of management-union relations at the enterprise level based on mutual trust (Shukan Rodo News, May 22, 1995; Gekkan Rodo Kumiai, 1995: 68-73). Observers and union leaders comment that the policies proposed by the Nikkeiren document are much more drastic ones than those proposed by previous documents of the employers' association in the past twenty years because they see the document as representing employers' attempts to considerably weaken employment security and to virtually do away with seniority wages (e.g., Kyotani, 1995; Nakanishi, 1995; Shukan Rodo News, September 11, 1995).(7) As stated above, employment security and seniority wages are two of the three components of "Japanese-style employment relations." It should be noted that the employers' association intends to retain the third component of the employment relations, i.e., harmonious relations between employers and unions at the enterprise level, in spite of the drastic reform of the other two components. In other words, employers expect enterprise unions (at least some of them) to cooperate with the radical reform of employment relations and are thus calling for "cross-class alliances" with them. As shown below, responses of unions to the Nikkeiren document were divided between opposition and lukewarm support.

The restructuring of "Japanese-style employment relations" entails the deregulation of labor relations through the revision of the Labor Standards Law and other labor-related laws. The deregulation will make it much easier for employers to restructure employment relations in line with the policy recommendations of Nikkeiren. The government formally decided to make employment relations a target of deregulation and established an advisory council to the Minister of Labor to review regulations on labor relations under the current Labor Standards Law. In December 1997, the council recommended, among other things, the expansion of work categories eligible for the discretionary work scheme, the extension of the upper limit of fixed-term labor contracts from one to three years, and the regulation of overtime work hours based on a standard guideline set by the Minister of Labor, rather than a legally binding upper limit (Araki, 1997; Japan Labor Bulletin, February 1, 1998). The government drafted a bill for the revision of the Labor Standards Law based on the recommendations and submitted it to the ordinary session of the National Diet in April, 1998. Nikkeiren takes the position that the Labor Standards Law should not decide details of how each worker works and highly evaluates the proposed revision of the law as "the first step in the right direction" (Shukan Rodo News, February 16, 1998a). Rengo and the two other confederations, on the other hand, oppose the proposed revision. For example, Rengo argues that the deregulation of employment relations will promote "the individualization of labor contracts" that not only lowers the level of working conditions, but also poses a threat to the very existence of labor unions (Shukan Rodo News, February 16, 1998b). Because of the disagreements of the concerned parties, the Diet did not pass the bill and decided to carry the bill over to the next session.

The economic recession put pressure on the government not only to deregulate its control on employment relations, but also to restructure the government itself into a more efficient form. The restructuring of the government, commonly called "administrative reform," can be seen as another form of deregulation. Although the administrative reform started in the 1980s with the privatization of telecommunication and railway services, it did not proceed smoothly thereafter because plans to privatize or streamline government services met opposition from actors who had vested interests in the existing government structure. The severe recession in the 1990s has revitalized political debates over administrative reform. The issue most relevant to the internal politics of the labor movement is the political debate over the privatization of postal services. This is because such an issue creates a division between public-sector unions opposed to a drastic reform of the government and private-sector unions, particularly those in export-oriented industries, in support of a small and efficient government (see below).

This section has shown that economic conditions have been unfavorable for the labor movement since the formation of Rengo. Particularly important for the labor movement is the fact that employers have started to withdraw their support for employment security and seniority due to the inflexibility of the existing system of employment relations. Since these two components of employment relations are an integral part of the industrial relations institutions in the private sector since the late 1960s, the withdrawal of employers' support suggests that institutional arrangements of industrial relations may go through a fundamental transformation.8 If so, unions have to re-evaluate the policies that they have taken so far and to make policy choices in response to the transformation of institutional arrangements.

(1) Sectoral Politics within Rengo(9)

The changes in the context of the labor movement, discussed in the previous section, have sparked inter-union debates within Rengo over union policies. Since its establishment in 1989, Rengo has held five conventions. Observers comment that the first three conventions (held in 1989, 1991, and 1993) were devoid of active debates and largely ceremonial (Shimada and Inoue, 1994: 10-11; Egami, 1997: 8). However, delegates engaged in active debates in the conventions in 1995 and 1997 because some of them became frustrated with the ineffectiveness of the confederation in dealing with problems confronting the labor movement. Active debates surprised many observers of Rengo because they had regarded the internal politics of Rengo as being largely de-politicized under the dominance of big business unions.

In the 1995 convention, the executive committee of Rengo proposed a program that was widely seen as reflecting policies of big business unions. The program stated that the Japanese society entered into a "mature stage" characterized by slow economic growth, and that unions should strive to establish "the system of consensus building at every level of society," particularly at the macro- and society-wide-level. It emphasized the macro-level participation of unions because enterprise-level activities of unions were not sufficient to improve living standards of workers under slow economic growth and because macro-level activities of unions were necessary to ensure fair distribution of resources at the societal level (Shukan Rodo News, September 18, 1995; October 9, 1995a, b). In other words, the program reaffirmed the position taken by big business unions in the unification movement in the 1980s, particularly the neo-corporatist-style policy participation at the national level. Delegates of unions representing workers in small- and medium-sized firms, however, expressed dissatisfaction with the "mature society thesis" of the program. They criticized the "mature society thesis" for being too vague to be a guideline of union policies and argued that the thesis indicates the halfheartedness of Rengo in addressing issues that workers in small- and medium-sized firms face, such as improvement of working conditions. They also expressed frustration at what they saw as meager results of Rengo's activities in the previous six years in the areas of wage increases, employment security, and political issues (Shukan Rodo News, October 9, 1995a, b). In this way, policy debates in the 1995 convention mainly reflected the difference in interests between big business unions that were relatively satisfied with the status quo of industrial relations and unions of small- and medium-sized firms that were strongly dissatisfied with it.

Debates in the 1997 convention reflected conflicts not only between unions of different firm sizes, but also between unions in the private and public sectors. In contrast to the program of the previous convention, the tone of the program adopted by the convention was critical. The program stated that the restructuring of employment relations and the revision of the Labor Standards Law were "the manifestation of the supremacy of market principle" and warned that the supremacy of market principle would increase social inequality and deprive society of its vitality (Shukan Rodo News, September 22, 1997). The critical tone of the program seemed to reflect the criticism of many delegates to the previous convention against the acquiescent tone of "the mature society thesis" as well as a sense of crisis shared by Rengo's leaders that the deregulation of employment relations would seriously undermine the organizational base of the labor movement (ibid.).

Debates in the convention reflected inter-union politics based on sectoral interests. They centered on the issues of the deregulation of employment relations and administrative reform, particularly the privatization of postal services. While delegates of unions of small- and medium-sized firms(10) and public-sector unions actively engaged in discussions about these two issues, those of unions representing workers in large private-sector firms largely remained silent (Shukan Rodo News, October 6, 1997; Matsubara, et al., 1997: 28-29). The "silence" of big business unions means two things. First, big business unions do not necessarily oppose the revision of the Labor Standards Law. For example, Denki Rengo (Japanese Electrical, Electronic and Information Union) and Jidosha Soren (Confederation of Japan Automobile Workers' Unions) support the expansion of work categories eligible for the discretionary work scheme, one of the main points in the proposed revision of the law (Shukan Rodo News, February 16, 1998b; April 13, 1998). Unions representing workers in small- and medium-sized firms, on the other hand, strongly oppose the revision of the law because they fear that the revised law will make relatively poor working conditions in these firms even poorer (see Rodo Undo, 1998: 65-6; also see below). Second, public-sector unions and big business unions disagree on the issue of administrative reform. Although public-sector unions accept the necessity of making government organizations more efficient, they oppose the privatization of postal services and drastic restructuring of other government organs. Big business unions, on the other hand, see the big and inefficient government as hampering the competitiveness of export-oriented firms and demand a drastic form of administrative reform. For example, Kinzoku Rokyo (Japanese Council of International Metalworkers' Unions or IMF-JC, a loose federation of unions in export-oriented metal working industries, such as steel, automobile, electronics, and shipbuilding) adopted, in its convention held just before the 1997 convention of Rengo, a resolution calling for a thorough implementation of administrative reform including the privatization of postal services in order to reduce expenditures of the government (Shukan Rodo News, September 8, 1997). The pro-administrative reform position of IMF-JC is said to have created tension between big business unions and public-sector unions such as Zentei (Japan Postal Workers' Union) and Jichiro (All-Japan Prefectural and Municipal Workers' Union) (Matsubara, et al., 1997: 35).

Big business unions have also taken much more positive attitudes toward the restructuring of "Japanese-style employment relations," as envisioned by the 1995 Nikkeiren document, than unions in other sectors. For example, Denki Rengo supports the shift from seniority-based wages to merit-based wages on the condition that the employment security of union members is maintained. Its position is said to reflect the interests of high-skilled white collar workers such as engineers and researchers. Shogyo Roren (Japanese Federation of Commercial Workers' Unions) proposes the division of employees into the "stock" and "flow" types as well as merit-based pay. The proposal of Shogyo Roren is very similar to that proposed by Nikkeiren. Besides the unions of electronic and commercial workers, those of workers in petroleum, chemical, automobile, and telecommunication industries show partial, if not full, support for Nikkeiren's policy on the restructuring of employment relations (Shukan Rodo News, September 11, 1995; Kijima, 1997; Matsunaga, 1998). In this way, these unions respond more or less positively to the employers' call for "cross-class alliances" (see above).

Some big business unions have gone further to form alliances with employers in a more explicit way. Unions in export-oriented industries formed "cross-class alliances" against unions in industries sheltered from the pressure of international competition. For example, IMF-JC, consisting of unions in automobile, electronics, steel, and shipbuilding industries, issued a joint statement with Nikkeiren, demanding the government to speed up deregulation and administrative reform (Shinoda, 1995). The formation of such an alliance was facilitated by wage politics between the export-oriented and sheltered sectors. Unions in export-oriented industries have become increasingly frustrated with the existing system of annual rounds of wage negotiations (so called Shunto or spring offensive) because the system has tended to reduce wage differentials between high-productivity and low-productivity industries(11) and allowed unions in the latter industries to "free ride" wages decided by those in the former industries.(12) Thus, these unions have demanded that a wage level of each industry be determined by productivity of that industry, and that regulations be removed from the hitherto protected industries (Shinoda, 1995; Hyodo, 1997: 523). The formation of the "cross-class alliance" between IMF-JC and Nikkeiren reflected the convergence of interests between big business unions and employers in the "deregulation" of the system of wage determination.(13)

In this way, the inter-union politics within Rengo have become active since the mid-1990s, as sectoral interests have generated diverse responses to the transformation of industrial relations among industry-level federations affiliated with the confederation. While unions of large private-sector firms have largely accepted the logic of market competition and supported deregulation and administrative reform, unions of small-and medium-sized private-sector firms and public-sector unions have been reluctant to accept the logic of market competition and opposed the deregulation of employment relations and the drastic restructuring of the government. Why have sectoral-based inter-union politics come to the fore in this period? One of the main reasons is the severe recession in the 1990s. As Peter Swenson points out in the case of labor politics in Sweden:

Economic stagnation or a crisis of competitiveness...can turn latent conflicts within labor into open rivalry, making private-sector unions available for external alliances [i.e., cross-class alliances]. Zero-sum conflict among unions over shares of national income can disrupt previously stable relations within their confederations (Swenson, 1992: 50).

The labor movement in Japan seems to have experienced similar dynamics of inter-union politics. While conflicts within labor developed mainly between unions in export-oriented industries and public sector-unions in Sweden, such conflicts in Japan have developed not only between these two groups of unions, but between big business unions and unions of small- and medium-sized firms. It can be argued that the dualist structure of labor markets has contributed to the latter type of conflicts within labor. Labor markets in Japan have been characterized by dualism because of differences in wages and working conditions between workers in large firms and those in small- and medium-sized firms. For example, workers in manufacturing firms with less than 99 employees and with 100-999 employees received, respectively, 71.5 percent and 79.3 percent of wages earned by workers in firms with more than 1,000 employees in 1995. Concerning working conditions, while 80 percent of large-sized firms have adopted the five-day-week system, it is introduced to only 40 percent of medium-sized firms and 33 percent of small-sized firms (Igarashi, 1998: 270). The dualist structure of labor markets is also indicated by the uneven distribution of union density among firms of different sizes: while 59.3 percent of workers employed by large-sized firms (more than 1,000 employees) are organized by unions, the density of workers employed by medium-sized firms (100-999 employees) and small-sized firms (less than 100 employees) is 25.0 percent and 4.1 percent, respectively (ibid.). Active inter-union politics between big business unions and unions of small- and medium-sized firms seem to reflect the intensification of dualism due to the severe recession in the 1990s.

(2) Internal Politics within Big Business Unions

So far, I have examined inter-union politics within the labor movement based on sectoral interests. An implicit assumption in the sectoral-based analysis is that unions act as unitary actors because their behaviors are shaped by conditions of sectors where unions organize their members. However, this assumption has some limits because unions themselves consist of top leaders, lower-ranking officials, and rank and file members and may have political dynamics within them.(14) In this sub-section, I examine a "rank and file movement" within Denki Rengo, the industry-level federation of enterprise unions in the electronics industry, and point out the limits of the movement in challenging top union leaders. The examination has to be brief because data on the internal politics of the union are sketchy.

Denki Rengo consists of enterprise unions of big electronics firms such as NEC, Fujitsu, Hitachi, and Toshiba. These enterprise unions actively cooperate with employers in their efforts to make employment relations flexible, such as the introduction of the discretionary work scheme (e.g., Asahi Shinbun, March 3, 1998). However, there are groups of activists critical of cooperative policies taken by union leaders within the enterprise unions of the major electronics firms. In 1988, these groups formed an industry-level body called Denki Rodosha Kondankai (the Discussion Meeting of Electronic Workers) in order to facilitate discussion and exchanges of information among activists in different enterprises. Activists characterize the enterprise unions as being controlled by management and criticize union leaders for failing to address the problems that many rank and file members face under the restructuring of employment relations. Such problems include low wage increases, the anxiety felt by union members about the future of their employment security, long work hours, and gender discrimination in wages. They "expose" these problems by conducting informal surveys among union members. These informal surveys provide opportunities for them to express their dissatisfactions with union leaders and their policies. According to activists, many rank and file members do not feel free to openly express their dissatisfactions through the formal channel of the unions because the unions are under the oligarchical rule of the pro-management leadership (Nakayama, 1997; Chingin To Shakai Hosho, 1995).

Formal surveys conducted by Denki Rengo in 1989 and 1994 show that about 20 percent of electronics workers are dissatisfied with their unions (Chosa Jiho, 1995).(15) If activists critical of union leaders can turn these dissatisfied workers into their support-base, their activities will pose a threat to the union leadership. However, the capacity of the "rank and file movement" to challenge the union leadership is seriously limited for two reasons. First, studies of "rank and file movements" in unions in the United States show that such movements may have an impact on the internal politics of unions only when activists involved in them establish their "stronghold" in union locals by winning union elections at the local level (e.g., Nyden, 1984). The "rank and file movement" in the Japanese electronics industry also enter candidates in union elections at the local level, but these candidates have a very slight chance of winning. This is because union officials make every effort, even at the cost of union democracy, to prevent candidates critical of cooperative union-management relations from winning elections. The second limit concerns the political orientation of the "rank and file movement" of electronics workers. Even if the groups of activists at the enterprise level and their industry-level body emphasize that their activities are politically neutral, it is widely believed that leaders of these groups are members or sympathizers of the JCP (Japan Communist Party). Since the support-base of the JCP among workers in large private-sector firms is very narrow,(16) the opposition movement led by communists within the enterprise unions of the major electronics firms is not expected to win widespread support among union members.

In this paper, I have examined the internal politics of the labor movement since the formation of Rengo. I show that the internal politics of Rengo became active as institutional arrangements of industrial relations went through a transformation due to the withdrawal of employers' support for "Japanese-style employment relations." The internal politics within the confederation reflect conflicts of interests between private-sector and public-sector unions as well as between big business unions and unions of small- and medium-sized firms in the private sector. Big business unions in principle support the deregulation of employment relations and other measures to make Japanese firms competitive in global markets, while public-sector unions and unions of small- and medium-sized firms oppose the drastic restructuring of employment relations. I pointed out the intensification of dualism in labor markets due to the economic recession as one of the background factors for the active internal politics within Rengo. In addition to the sector-based politics between unions, I have shown that unions at the industry- or enterprise-levels are not necessarily unitary actors and have their own internal political dynamics, as indicated by the existence of a "rank and file movement" within the electronics workers' unions.

What does the active inter-union politics indicate for the future of the labor movement in Japan? Will the labor movement go through fundamental changes? Is it going to move away from enterprise-based, largely harmonious relations with employers? It seems that the labor movement is going to be polarized between big business unions which represent "core" workers and unions representing "peripheral" workers. Workers in the "peripheral" sector include those who work at large private-sector firms but without employment security and those who work at small- and medium-sized firms. While the former unions are likely to maintain enterprise-based, cooperative relations with employers, the latter unions may move away from enterprise-based unionism to industry- or region-based unionism. Some unions of small- and medium-sized firms have attempted to organize the vast majority of unorganized workers in these firms on the basis of industry or region, rather than enterprise. These unions are likely to intensify their organizing efforts on the industrial or regional basis.(17) Furthermore, new types of regional unions have emerged with the aim of protecting specific types of workers, such as middle-aged workers in managerial positions most vulnerable to dismissal or forced retirement, and female workers who suffer from the widespread gender-based discrimination in wages and working conditions. When these regional- and industry-based unions become more influential than enterprise-based big business unions within Rengo and the labor movement in general, the labor movement in Japan will be transformed in a fundamental way. If big business unions continue to dominate Rengo, on the other hand, the influence of organized labor will decline further. This is because members of big business unions are limited to the ever shrinking "core" workers, and because these unions are not very interested in organizing new workers outside enterprises.

Araki, Takashi (1997): "Changing Japanese Labor Law in Light of Deregulation Drives: a Comparative Analysis." Japan Labor Bulletin, Vol.36, No.5.

Asahi Shinbun (March 3, 1998): "'Kobetsu ni Hataraki wo Hyoka' Kasoku." [the Acceleration of Individual Evaluation of Work]

Chingin to Shakai Hosho (1995): "Rodo Undo wo Omoshiroku Suru Ho." [How to make the Labor Movement More Interesting] Chingin to Shakai Hosho, No.1163, pp.4-31.

Chosa Jiho (1995): "Dai 12 Kai Kumiaiin Ishiki Chosa Ketsukka Hokoku." [Report on the Twelfth Opinion Survey of Union Members] Chosa Jiho, No.276.

----------.(1998): "Denki Rengo Seiji Anketo Chosa Ketsukka Hokoku Sho." [Report on Denki Rengo Political Survey] Chosa Jiho, No.302.

Egami, Sumio (1997): "Rengo Kettsei Hachi Nen, Atarashii Undo no Ibuki." [the Eighth Year of Rengo, a Sign of New Movement] Rodo Keizai Junpo, No.1599, pp.4-10.

Gekkan Rodo Kumiai (1995): "'Shin Nihonteki Keiei Shisutemu tou Kenkyu Purojekuto' Houkoku no Gaiyo." [the Outline of "New Japanese-style Management System Research Project" Report] Gekkan Rodo Kumiai, (October, 1995), pp.68-73.

Hyodo, Tsutomu (1997): Rodo no Sengoshi (ge). [the Postwar History of Labor (part 2)] Tokyo: university of Tokyo Press.

Igarashi, Jin (1998): Seito Seiji to Rodo Kumiai Undo. [Party Politics and the Labor Union Movement] Tokyo: Ochanomizu Shobo.

Inoue, Masao (1997): Shakai Henyo to Rodo. [the social transformation and Rengo] Tokyo: Bokutaku Sha.

Iversen, Torben (1996): "Power, Flexibility, and the Breakdown of Centralized Wage Bargaining: Denmark and Sweden in Comparative Perspective." Comparative Politics. Vol.28, pp.399-436.

Japan Labor Bulletin (1998): "Council Proposes Overhaul of the Labour Standards Law." Japan Labor Bulletin, Vol.37, No.02.

Kijima, Atsuo (1997): "Bunki, Konmei to Aratana Kizashi." [Division, Confuse, and the Sign of Change] Rodo Joho, No.491, pp.8-11.

Kyotani, Eiji (1995): "Nihonteki Roshi Kankei no Chikaku Hendo." [the Diastrophism of Japanese-style Industrial Relations] Chingin to Shakai Hosho, No.1159, pp.4-11.

Matsubara, Hiroshi, et al. (1997): "Zadankai Nihon Rodo Undo to Rengo." [A Roundtable Discussion on the Japanese Labor Movement and Rengo] Kokusai Rodo Undo (December, 1997), pp.25-40.

Matsunaga, Yukikata (1998): "Rengo no Henka no Kizashi to Rodo Undo." [the Sign of Rengo's Change and the Labor Movement] Shakai Shugi, No.417, pp.26-71.

Moody, Kim (1997): Workers in a Lean World: Unions in the International Economy. London: Verso.

Nakanishi, Satoru (1995): "'Han Rodosha Sengen' Nikkeiren Shin Senryaku no Motarasu Mono." ['the Anti-Worker Declaration' What the new strategy of Nikkeiren will bring] Gekkan Rodo Kumiai, (October, 1995), pp.12-15.

Nakayama, Morio (1997): "Shokuba no Henka ni Kakushin Fukameru." [We are Confident that Workplaces are Changing] Rodo Undo, No.384, pp.138-143.

Nihon Rodo Kenkyu Kiko (1997): Rodo Undo Hakusho. (Heisei 9nendo Ban). [White Paper of the Labor Movement (the 1997 edition)] Tokyo: Nihon Rodo Kenkyu Kiko.

Nikkeiren Times (November 26, 1987): "'Rengo' Hossoku wo Kangei." [We welcome the establishment of Rengo]

----------.(November 23, 1989): "Shin 'Rengo' Hossoku wo Kangei." [We welcome the establishment of New Rengo]

Nyden, Philip W. (1984): Steel Workers Rank-And-File: The Political Economy of a Union Reform Movement. Praeger.

Pontusson, Jonas (1995): "From Comparative Public Policy to Political Economy: Putting Political Institutions in Their Place and Taking Interests Seriously." Comparative Political Studies, (28) 1: 117-147.

Pontusson, Jonas and Swenson, Peter (1996): "Labor Markets, Production Strategies, and Wage Bargaining Institutions: the Swedish Employer Offensive in Comparative Perspective." Comparative Political Studies, (29) 2: 223-250.

Rodo Undo (1998): " 'Rengo' no Henka to Hirogaru Kyodo no Joken." [the Changes of Rengo and Increased Opportunities for Cooperation] Rodo Undo, No.395, pp.62-71.

Rodo Keizai Junpo (1994): "Kikai Kinzoku Soshiki no Toitsu ni Mukete."[Toward the Unification of Machine Metal Organizations] Rodo Keizai Junpo, No.1518, pp.21-27.

Saeki, Tetsuro (1992): "Zenroren no Soshiki to Undo." [the Organization and Movement of Zenroren] in Hosei Daigaku Ohara Shakai Mondai Kenkyu Jo (ed.) "Rengo Jidai" no Rodo Undo. Tokyo: Sogo Rodo Kenkyu Jo, pp.135-152.

Sato, Hiroki (1996): "Keeping Employees Employed: Shukko and Tenseki Job Transfers- Formation of a Labor Market within Corporate Groups." Japan Labor Bulletin, December 1, 1996.

Shimada, Kazuo and Inoue, Masao (1994): "Rengo Rodo Undo no Totatsuten to Kanosei." [What the Labor Movement led by Rengo Has Achieved and its Future Possibility] Rodo Keizai Junpo, No. 1518, pp.10-20.

Shinoda, Toru (1995): " 'The Tale of Cain and Abel?'- A Study of the Contemporary Japanese Labor Politics." Japan Labor Bulletin, Vol.34, No.11.

----------.(1996): "Futatabi 'Niwatori kara Ahirue?- 55 Nen Taisei no Hokai to Rengo" ["From Chicken to Duck" again?: the Collapse of the 1955 System and Rengo] in Nihon Seiji Gakkai (ed.) 55nen Taisei no Hokai. Tokyo: Iwanami Shoten.

----------.(1997): "Rengo and policy participation: Japanese-style neo-corporatism?" in Sako, Mari and Hiroki Sato (eds.), Japanese Labour and Management in Transition: Diversity, flexibility and participation. London: Routledge pp.187-214.

Shukan Rodo News (May 22, 1995): "Nikkeiren Houkoku, Teki Shokyu ha Minaoshi." [Nikkeiren Report, the Reconsideration of Regular Wage Raises]

----------.(September 11, 1995): "Yureru Nihongata Koyo Shisutemu." [Shaking Japanese-style Employment System]

----------.(September 18, 1995): "Rengo Shin Undo Hoshin An." [the Proposed New Policy Program of Rengo]

----------.(October 9, 1995a): "Rengo Taikai Undo Hoko ni Iradachi no Koe." [Irritated Voices Were Raised Against the Policy Program at Rengo Convention]

----------.(October 9, 1995b): "Undoron Kihaku ni Shosokan." [the Feeling of Impatience at the Lack of Philosophy]

----------.(September 8, 1997): "Kancho Saihen ya Yusei Jigyo Minaoshi wo Kyocho." [the Restructuring of Government Offices and the Reappraisal of Postal Services were Emphasized]

----------.(September 22, 1997): "Nihonteki Koyo Shisutemu no Henyo ni Yureru Roso." [Labor Unions are Shaken by the Transformation of Japanese Employment System]

----------.(October 6, 1997): "Rengo Shintaisei Nenmatsu ni Kakete Shiren." [the New Leadership of Rengo Will Face a Tough Time Toward the End of this Year]

----------.(February 16, 1998a): "Roki Ho Ha Hitsuteki Tenkan Wo." [the Labor Standards Law is Expected to Change Qualitatively]

----------.(February 16, 1998b): "Rengo Taian Wo Motoni Shusei Nerau." [Rengo Aims at the Revision Based on its Counterproposal]

----------.(April 13, 1998): "Rodo Kijun Ho Kaisei, Kokkai Shingi Chikazuku." [the Diet Deliberation on the Revision of the Labor Standards Law is Around the Corner]

Suzuki, Akira (1998): "Toward an Alternative Analytic Approach to the Internal Politics of Unions." International Journal of Japanese Sociology, No.7, pp.45-63.

Swenson, Peter (1991): "Bringing Capital Back In, or Social Democracy Reconsidered: Employer Power, Cross-Class Alliances, and Centralization of Industrial Relations in Denmark and Sweden." World Politics, 43, pp.513-44.

----------.(1992): "Union Politics, the Welfare State, and Intraclass Conflict in Sweden and Germany," in Golden, Miriam and Jonas Pontusson (eds.), Bargaining for Change: Union Politics in North America and Europe. Ithaca: Cornell University Press, pp.45-76.

Turner, Lowell (1992): "Industrial Relations and the Reorganization of Work in West Germany: Lessons for the U.S.," in Mishel, L. and Voos, P.B. (eds.), Union and Economic Competitiveness. New York: M.E. Sharpe, Inc., pp. 217-246.

Tsuru, Tsuyoshi and Rebitzer, James (1995): "The Limits of Enterprise Unionism: Prospects for Continuing Union Decline in Japan." British Journal of Industrial Relations, 33: 3, 459-492.

Yates, Charlotte A.B. (1993): From Plant to Politics: the Autoworkers Unions in Postwar Canada. Philadelphia: Temple University Press.

_______________________________

1 The spread of cooperative economic unionism among private sector unions was related to the decentralized structure of the labor movement. The decision-making authority of unions was decentralized at the enterprise level. Although many enterprise unions were affiliated (through industry-level federations of enterprise unions) with either one of the four national confederations, the confederations did not have much authority over enterprise unions. Since enterprise unions were relatively independent from the control of the confederations, management was able to influence the internal politics of these unions in such a way that the policy orientation of

unions would become cooperative with management. Thus, cooperative economic unionism spread not only among enterprise unions affiliated with Domei, but also among those affiliated with Sohyo.

2 While unions affiliated with Rengo are spread across major industries in the private as well as public sectors, unions affiliated with Zenroren and Zenrokyo are concentrated in specific industries and sector. Zenroren mainly organizes public sector workers (e.g., local government workers and teachers) and medical workers such as doctors and nurses. Zenrokyo mainly organizes workers in Japan Railways (JRs, privatized railway companies) and workers in Tokyo Metropolitan Government (Shinoda, 1997: 199).

3 In this section, I will examine changes in the economic context of the labor movement and will not deal with changes in the political context due to the limited space of the paper. The political context of the labor movement has changed drastically since the early 1990s (the realignment of political party started after the collapse of the dominance of the Liberal Democratic Party). However, it seems that economic context has a stronger impact on the internal politics of the labor movement than political context.

4 It should be noted that "Japanese-style employment relations" mainly covers male workers in larger firms. Most female workers and workers in smaller firms do not enjoy employment relations as secure and seniority-based as workers in larger firms.

5 Japanese firms coped with the recessions in the 1970s and 1980s, first through intra-firm transfers of employees, and then through "loaning" employees to outside firms. There are two ways of loaning workers: "shukko" and "tenseki." While "shukko" means that workers who are transferred to an outside firm retain their status as employees of their original firm, "tenseki" means those who are on loan no longer retain such status and become employees of receiving firms. Firms that receive loaned workers are in most cases subsidiaries and/or suppliers of sending firms and are smaller than the latter (see Sato, 1996).

6 Nikkeiren proposed that employers divide regular employees into three groups: (1) the group of employees whose skills are accumulated on a long-term basis, (2) the group of employees with specialized skills, and (3) the group of employees with general skills, and that the three groups be employed on long-, medium-, and short-term basis, respectively. It also proposed that each employer develop its own "portfolio of the three groups of employees" so as to enhance the competitiveness of the firm (Shukan Rodo News, May 22, 1995; Gekkan Rodo Kumiai, 1995: 68-73).

7 Nikkeiren proposed a guideline of personnel policies to its member firms, and the member firms do not necessarily implement policies as proposed by the employers' association. However, it should be noted that policy proposals of Nikkeiren are widely seen as representing the general interests of employers and have a strong impact on the discourse of employment relations. It should also be noted that the Nikkeiren document of 1995 cites cases of 14 firms whose personnel policies are close to those proposed by the employers' association.

8 Pontusson and Swenson argue that employers are the principal agents of institutional change. According to them, the transformation of industrial relations institutions in Sweden from the nationally centralized collective bargaining to the industry- or enterprise-based decentralized bargaining occurred largely because employers in export-oriented, engineering industries no longer see the centralized bargaining as worth supporting. The centralized bargaining in Sweden had been institutionalized in the 1930s in order to control wage militancy of construction workers. However, in later years, the centralized bargaining failed to restrain wage increases of workers sheltered from international competition, and the export-oriented employers began to see the centralized bargaining as being rigid and burdensome (see, Pontusson and Swenson, 1996; Swenson, 1991).

9 By "sector," I mean a group of industries that share similar characteristics (e.g., export-oriented sector). I also mean a group of firms with similar sizes (e.g., the sector of small- and medium-sized firms).

10 The unions that most actively engaged in discussions are Zensen Domei (the Japanese Federation of Textile, Garment, Chemical, Mercantile, Food and Allied Industries Workers' Unions), Zenkoku Ippan (National Union of General Workers), and Kinzoku Kikai (National Metal and Machinery Workers' Unions of Japan).

11 Readers may wonder why the wage leveling occurred under the decentralized system of industrial relations. Although collective bargaining is decentralized at the enterprise level, wage settlements have been coordinated within and across industries by the system of annual rounds of wage negotiations (Shunto).

12 According to high-productivity industry unions, the "free-riding" have lowered the rate of wage increases in export-oriented industries and contributed to higher prices of products and services of protected industries. Export-oriented employers have used the possible spill-over of wage increases into the high-productivity sector into the low-productivity sector as an excuse for offering wage increases much lower than demanded by the unions. In addition, employers in sheltered, low-productivity industries have shifted higher labor costs due to wage increases onto prices (Shinoda, 1995; Shinoda, 1996: 142; Hyodo, 1997: 523).

13 Note that the inter-union wage politics in the mid-1990s in Japan are not unlike wage politics in other industrialized countries in the 1980s. For example, unions representing skilled workers in the competitive sector in Denmark and Sweden were against wage-leveling under centralized collective bargaining and formed cross-class alliances with employers in order to decentralize the system of wage bargaining (see Iversen, 1996).

14 I review theories of the internal politics of unions elsewhere (see Suzuki, 1998).

15 This does not mean that the rest of respondents said they were satisfied with their unions. In both surveys, about 45 percent of them said that they could not say either satisfied or dissatisfied, and about 30 percent of them said they were satisfied (Chosa Jiho, 1995).

16 A survey of workers in the electronics industry in 1997 shows that only 1.8 percent of respondents said they support the JCP (Chosa Jiho, 1998).

17 This is indicated by the amalgamation of two unions organizing workers in small- and medium-sized firms in machine tool and other metal-working industries, Kinzoku Kikai (National Metal and Machinery Workers' Unions of Japan) and Zenkin-Rengo (Japanese Federation of Metal Industry Workers' Unions). Prior to the establishment of Rengo, these two unions had been affiliated with Sohyo and Domei, respectively, and engaged in intense inter-union competition. However, the rivalry between the two unions has become less intense since the formation of Rengo. Thus, the two unions have started discussions about amalgamation so that the amalgamated union will be able to unionize unorganized workers more effectively by pooling their resources (Rodo Keizai Junpo, 1994: 21-24).