From Class Struggle to General Middle-Class Society to Divided Society: Societal Models of Inequality in Postwar Japan

Transcription

1 Social Science Japan Journal Vol. 11, No. 1, pp Published online May 13, 2008 doi: /ssjj/jyn From Class Struggle to General Middle-Class Society to Divided Society: Societal Models of Inequality in Postwar Japan David CHIAVACCI* After the fierce class struggles in the first postwar years, a societal model describing Japan as a general middle-class society with outstanding equality in opportunities and outcome became dominant. In recent years, a new societal model of Japan as a divided society has replaced this general middle-class model. Nonetheless, empirical research and comparative studies neither fully support a model of Japan as an exceptionally equal society from the 1960s onward nor do they show a fundamental transformation of contemporary Japan into a socially divided society. This paper argues that the sequence and timing of societal models of inequality in Japan since 1945 reflect the degree of resonance that societal models of inequality have in the lifeworlds of society. 1. Introduction Social class and social stratification is one of the basic and founding concepts of sociology. Moreover, questions concerning inequality and its desirable degree are a continuous and fundamental political issue of modern society. Therefore, research on social stratification, income distribution or social mobility occupies not only a central position in the social sciences but it is also ideologically contested like maybe no other research field. Which discourse and perception of inequality becomes dominant and is generalized into societal models of inequality is not only a question of scientific truth but also of political struggle. Postwar Japan is a very good case study to analyze the making and unmaking of societal models of inequality because a sequence of clearly opposite models can be identified. During the early postwar years, the perception of social inequality was dominated in Japan by class struggle and class division. However, from the mid-1960s onward, Japan was increasingly regarded as a general middle-class society. This view was further consolidated and enhanced during the 1970s. According to foreign descriptions as well as a dominant Japanese self-view, Japan was an exceptionally equal society regarding chances and outcome in international comparison. Japan had, according to this societal model, a very equal income distribution per household and was a society with an outstanding degree of social openness. Thanks to Japan s highly meritocratic school system, it was... David CHIAVACCI is a Research Fellow and Reader at the Institute of East Asian Studies, Free University Berlin. His recent publications include The Social Basis of Developmental Capitalism in Japan: From Postwar Mobilization to Current Stress Symptoms and Future Disintegration in Asian Business & Management 6(1): (2007) and The General Middle Class Model under Pressure: Mainstream Japan at a Turning Point? in Hede Helfrich, Melanie Zillekens and Erich Hölter, eds., Culture and Development in Japan and Germany (Daedalus 2006). He can be reached at the Institute of East Asian Studies, Free University Berlin, Ehrenbergstrasse 26-28, D Berlin, Germany, or by at dachiava@zedat.fu-berlin.de. *The author would like to thank Verena Blechinger-Talcott, Wolfram Manzenreiter, Iris Wieczorek and three anonymous reviewers of Social Science Japan Journal for their substantial comments regarding earlier drafts of this paper. ª The Author Published by Oxford University Press in conjunction with the University of Tokyo. All rights reserved.

2 6 David CHIAVACCI believed that individual effort and not social origin are decisive for status attainment. Yet, from the late 1990s onward, the perception of Japanese society has changed completely, especially in Japan itself. Current public debate is marked by intensive discussion about rising inequality and new social disparities. Numerous contributions about these changes have been published in academia and the mass media. Japan is no longer described as a general middle-class society (sochuryu shakai), but as a divided society (kakusa shakai). How can the stark discrepancies of these societal models and their sequence and timing be explained? A simple answer would be that these models reflect the overall social structures of inequality and their transformation as reported by the social sciences. On the one hand, from the mid-1960s onward, Japan s income distribution had become exceptionally equal and its social stratification system had become exceptionally fluid concerning social mobility in international comparison. On the other hand, from the late 1990s onwards, Japan s income distribution has become very unequal and its social stratification system has become strongly closed, which means that social origin plays again a much more important role in educational attainments and career chances. Still, as will be discussed fully below, empirical research does not confirm such paradigmatic transformations of Japan s social structures of inequality. Overall, on the one hand, empirical support for a societal model of Japan as a general middle-class society from mid-1960s up to mid-1990s is very weak. Already before the late 1990s, Japan was neither an exemplary case of equality in outcome nor an exemplary case of equality in chances. However, on the other hand, empirical studies are also not clearly supporting a thesis of a complete transformation and radical new social differentiation in Japan in recent years. Based on available studies, it is hard to argue that contemporary Japan has suddenly been transformed into a divided society. Hence, empirical research results leave us quite at a loss concerning the sequence and timing of societal models of inequality in Japan. Does this mean that the establishment of societal models of inequality is all politics? Respectively, does this suggest that empirical findings are, at best, means in political agendas, selectively (mis)used by powerful political actors in order to establish a certain view of social inequality that suits their goals? Undoubtedly, the establishment of societal models of inequality is a political struggle. However, has this process been overwhelmingly dominated in postwar Japan by the so-called iron triangle or conservative establishment of the Liberal Democratic Party (LDP), bureaucrats and big business? As will be discussed below, the conservative establishment played a central role in the institutionalization of the general middle-class model and this model suited their political agenda. Still, they never completely dominated research and public debate concerning social inequality. Moreover, the new societal model of a divided society does not really serve the interests of the establishment. Although at least part of the conservative establishment played a pivotal part in preparing the ground for this new societal model of inequality, they are clearly on the defensive concerning this issue, and this new model is actually undermining their power position. This paper argues that to fully understand the sequence and timing of societal models of inequality in Japan, we need to bring society back in. The processes of establishing and displacing societal models of inequality are not simply a question of social structures or political interests, but are also due to the ability of such models to give meaning to the events and experiences of social inequality in the lifeworlds of broad segments of society. The establishment of a general middle-class society model and its recent displacement by a new model of a divided society is not simply reflecting a fundamental transformation of Japan s social structures and is not solely resulting from changing agendas of powerful political actors. These processes are also reflecting the degree of resonance that societal models of inequality have in the lifeworlds of Japan.

3 From Class Struggle to General Middle-Class Society to Divided Society 7 2. Production and Resonance of Societal Models of Inequality: A Framework In order to trace back the sequence and timing of societal models of inequality in postwar Japan, we need to analyze a two-fold process: (1) The production of societal models of inequality as an interaction between social science, mass media and political actors on the system or macro level, through which complex macrostructures of social inequality are translated and generalized into simplified and comprehensible models. (2) The degree of resonance of these models in the lifeworlds, 1 i.e. daily life at the micro level of society, through which these models may be able to attain a dominant position in society by being accepted by a majority as common sense concerning social structures of inequality. The analysis of these two processes leads us to one of the main (and overall still unresolved) enigmas in sociology: How are macro and micro, structure and agency, or system and lifeworld connected (cf. Fuchs 2001)? Or more concretely for this paper: How is the relation between societal models of inequality at the system level and daily experiences and social action at the level of lifeworlds? The theoretical framework of this paper for the production of societal models of inequality and their resonance in the lifeworlds draws from the insights and concepts of the scholarship about framing processes of social movements, 2 which has been developed and proliferated in recent years by building on Goffman s (1974) framework. Still, the production and resonance of societal models of inequality are in several aspects different from the framing processes of social movements. In order to clarify the theoretical framework, these differences are highlighted in the following discussion. 2.1 Production of Societal Models of Inequality at the System Level The production of societal models of inequality in modern society is a complex interplay between social science, mass media and powerful political actors. Intellectuals/social scientists create complex models of social inequality based on their empirical research, theoretical considerations and worldviews. Mass media and powerful political actors take up such complex models and disseminate them in a simplified form into public debate. Through this propagation, political actors try to establish a model that suites their own political agenda. As a result, societal models of inequality have always an ideological aspect and are often used by political actors to build or retain a hegemonic position. In contrast to the framing processes of social movements (cf. Benford and Snow 2000: ), this production process is not part of general social action or agency at the micro levels of the lifeworlds because of an insuperable gap between everyday experience of social inequality and the overall macrostructures of social inequality. Social actors in their lifeworlds have only a limited experience regarding the degree of income inequality and social openness of the whole of society. They are not able to construct directly a societal model of inequality from their own everyday experience. (In order to analyze macrostructures of social inequality, social scientists need to take a position as second-order observer beyond everyday experience and have to use constructed macro-indicators). Hence, the production Schütz and his disciples introduced the concept of lifeworld into sociology by adopting it from Husserl s phenomenology (e.g. Schütz and Luckmann 1973). In its original use, the concept was understood as the everyday world of social action of direct (i.e. prescientific) and unquestioned experiences. However, this paper is based on a different understanding of lifeworld. Habermas (1981) argues that modernization includes a colonization of lifeworld by the system. By adapting this view, this paper is based on the idea that modernization leads to an uprooting and fluidization of the everyday world and that societal models are needed to reestablish meaning and social order. Furthermore, as this paper is about social inequality, lifeworld is used in the plural because of significant differences in the lifeworlds of each social class. 2. For an overview, see Benford and Snow (2000).

4 8 David CHIAVACCI process is limited to a circle of social actors on the system level that have acquired specific knowledge like intellectuals and social scientists or occupy power positions like mass media or political actors. Overall, the production process results in strongly reduced and aggregated societal models of inequality that function as a generative frame for inequality and contain a number of central assumptions. The whole process of production is, in a democratic political system, so open and diversified that different societal models of inequality coexist and are in competition with each other. Whether one of these models, and which gains a dominant position, is a question of support and active dissemination by political actors and mass media. Still, it is also due to its degree of resonance in the lifeworlds. Societal models of inequality cannot simply be forced on social actors on the micro level. In order to attain a dominant position, societal models of inequality need to resonate among social actors by not contradicting their experiences in daily life. Clear contradictions between lifeworlds and model lead ultimately to a disconfirmation of a social model of inequality and to its loss of a dominant position Resonance of Societal Models of Inequality in the LifeWorlds Relative salience has been identified as an important factor for the degree of resonance of collective action frames of social movements (Benford and Snow 2000: ). However, societal models of inequality always have relative salience for social actors because inequality is continuously experienced in the lifeworlds. Moreover, the mode of resonance of societal models of inequality is qualitatively different from collective action frames of social movements. They do not have to mobilize social actors or gain their active support like collective action frames. In order to gain a dominant position, societal models of inequality need primarily to be passively accepted in the lifeworlds through their empirical credibility and the credibility of their articulators. Empirical credibility means that qualitative and quantitative differences in consumption patterns that are experienced in the lifeworlds are not contradicting the assumptions concerning income distribution of a societal model of inequality and that experienced patterns of social mobility are not contradicting the assumptions concerning the openness of the social stratification system. Empirical credibility of the articulators implies, especially in the case of political actors, that the societal model of inequality that they support is not contradicted by other policies and positions that the articulators promote. Furthermore, the resonance of a societal model of inequality also depends on its degree of embeddedness in the zeitgeist and general social development what has been called cultural resonance in collective actions frame research (Benford and Snow 2000: 622). In order to attain and retain a dominant position, a model needs to connect to the mainstream. This means also that the degree of resonance is not only due to the empirical accuracy of a model regarding events and experiences in the lifeworlds but also due to the ability of a model to give meaning to everyday practice in a changing world. If a societal model of inequality attains a dominant position, it has a crucial impact on social actors by shaping their perception of social inequality. It becomes part of the mainstream as a frame influencing and arranging everyday experience of inequality. It can even gain a prescriptive quality and channel social action into certain patterns. However, a societal model of inequality retains its dominant position only through its continuous resonance in the lifeworlds. Even a strongly enrooted and dominant societal model can still be contradicted through everyday experience, which leads For a study about a disconfirmation of a collective action frame due to its contradiction in everyday experience, see Babb (1996).

5 From Class Struggle to General Middle-Class Society to Divided Society 9 ultimately to disconfirmation. Overall, according to this theoretical framework, social reality is not based on a causative relationship between societal models and everyday experience, but on a constitutive relationship through which both factors are shaping and reinforcing each other in a constant loop mechanism. 3. Development and Timing of Societal Models of Inequality in Postwar Japan Four main junction points can be identified in the postwar development of societal models of inequality in Japan: (1) The failed attempt to establish a mass society model in the late 1950s. (2) The birth of the general middle-class model in the mid-1960s. (3) The enhancement of the general middle-class model during the 1970s. (4) The replacement of the general middle-class model by the divided society model from the late 1990s onward. Each of these four junction points in postwar development is analyzed based on the theoretical framework discussed above. Due to limitations of space, the focus is primarily on the resonance of societal models of inequality. 3.1 The First Debate: A Failed Try to Establish a Middle-Mass Society Model During the late 1950s, a first debate about social stratification developed between social scientists (cf. Kosaka 2000: ; Hashimoto 2003: 16 20). Young social scientists, including among others Matsushita Kei ichi and Fujita Shozo, identified a transformation of Japan and proposed a new model of social stratification. Adopting a non-marxist perspective of the US-American mass society theory, 4 they suggested a blurring of social stratification and the emergence of a large and dominant new middle class in Japan. This proposed model of Japan as a middle-mass society met unsurprisingly with a strong backlash from Marxist scholars, who were at the time an important fraction in Japanese academia. Marxist scholars fiercely criticized the proposal of their colleagues and emphasized the unaltered class character of Japan. They also showed that the new middle class, although expanding, was still a small minority. Overall, the debate ceased, without which the young social scientists could have prevailed in the dispute. A new model of Japan as a middle-mass society did not gain a dominant position. The failure of this proposed new societal model of inequality seems, on the one hand, rather surprising due to two changes, which should give credibility to this new model. First, a drastic deconcentration of income at the top occurred during World War II, which led to a much more equal income distribution (Moriguchi and Saez 2006). The main reason for this dramatic change was the transformation of Japan s political-economic system with the introduction of the 1940 system (Noguchi 1995) and the wealth destruction at the end of World War II. Postwar reforms like the land reform, tax system reform and the dissolution of the large zaibatsu made this significantly more equal level of In contrast to the discussion in the US, some proponents of a transformation of Japan into a mass society assessed this change more positively. In fact, in the US, the tenor was a sharp criticism of mass society as over harmonious and based on conformism (Thomson 2005: ). This critical description of the US at the time has strong similarities with criticism of Japan later on as a nonindividualistic society of sararıman clones (e.g. Sugahara 1994).

6 10 David CHIAVACCI income and wealth distribution permanent. Second, mass consumption had already begun to make some inroads into Japanese society during the second half of the 1950s (cf. Partner 1999: ). After the years of hardship during the war and the postwar reconstruction, daily life in Japan started to gain a still faint, but sweet taste of well-being beyond the daily struggle for survival. Furthermore, the idea of a Japanese middle-mass society attracted some powerful supporters (cf. Kosaka 2000: 149; Hashimoto 2003: 117). The newly formed Democratic Socialist Party as well as the LDP tried to exploit the idea of a new dominant middle class in Japan. However, on the other hand, times were characterizedbyopenpoliticalandsocialconflicts.class struggle and class division were still engraved in daily life and people s identity and any model of Japan as a classless middle-mass society was doomed by factual events. The struggles between newly organized and strong labor unions and management that had marked the first postwar years were still not overcome. On the contrary, the huge Mi ike strike of was a symbolic culmination of these labor conflicts that was accompanied by a solidarity movement in the whole country (Endoh 2000): The confrontation between the radical unions and the state culminated in the Great Miike Strike ( ), the longest and biggest strike in Japan s labor history. The Miike workers tried to mobilize a nationwide labor movement against capital and the state while gaining financial and moral support from the society. Labor Minister Hirohide Ishii admitted that this movement mobilized more people than the total who fought in the great Satsuma Rebellion early in the Meiji period. Simultaneously, a nationwide popular protest movement tried to prevent a renewal of the Treaty of Cooperation and Security with the US. It resulted in huge demonstrations in Tokyo and fierce clashes between protesters and riot police. Actual events presented a politically and socially divided Japan that was anything other than a harmonious middle-mass society. The voices of middle-mass society were simply drowned out in the turbulences of the late 1950s. 3.2 Birth of the General Middle-Class Model From the mid-1960s onward, over 90% of the Japanese population self-identified with a social middle position in the representative national surveys of the Cabinet Office (CAO) (Figure 1). This new trend was highlighted in the yearly Kokumin Seikatsu Hakusho (White Paper on the National Lifestyle) of the Japanese government. Mainstream mass media not only reported and stressed this increasing middle-class consciousness, 5 but journalists created slogans like 90% middle class (kyu wari churyu), all-encompassing middle class (oru chukan) or 100 million general middle class (ichioku sochuryu) and started to use them regularly. This reporting constitutes the birth of the general middle-class model. The main producers of this new discourse were not social scientists or intellectuals, but the mass media. At the time, the model was still very fuzzy. Japan was portrayed as a society with a nearly allencompassing middle-class consciousness. However, this assumed general middle-class consciousness, which was from the 1970s onward even regarded as distinctively Japanese, has a very weak empirical basis. The survey of the CAO provides five possible answer categories regarding the general lifestyle level, of which three are upper middle, middle middle and lower middle. It is not surprising that a large majority of the respondents identified themselves with one of the three middle categories. In fact, See, for example, how even the critical Asahi Shinbun (1964, 1967) adopted this line.

7 From Class Struggle to General Middle-Class Society to Divided Society 11 70% 60% 50% 40% 30% 20% 10% 0% Low Lower Middle Middle Middle Upper Middle Top Figure 1. Self-Identified Level of Lifestyle of Japanese Population, Source: CAO (2007). surveys with the same question and answer categories in other advanced industrial economies produce the same results of over 90% middle-class identification (Hayashi 1995: 53). Even in developing and semi-industrial countries, a large majority identifies itself with the middle class in similar opinion polls (Watanabe 1997: 57). Nonetheless, what is exceptional in the case of Japan is not that over 90% of the respondents identified themselves with one of the three middle positions from 1964 onward, but that clearly over 10% of the respondents identified in the surveys from 1958 to 1961 with a low level of lifestyle. 6 This is an indicator for the general impoverishment of Japanese society in the first postwar years due to the war. Self-identification as being part of the underprivileged was not at all stigmatized at the time. Wealth and income might have been much more equally distributed than in the prewar era, but even around 1960 it was still an equality of austerity and not an equality of an affluent lifestyle despite the first signs of mass consumerism from the late 1950s onwards. 7 The increasing resonance of this new model of Japan as a general middle-class society during the 1960s, despite its very shaky empiric foundation and fuzziness, was due to its timing in parallel with three significant changes. These changes transformed not only the overall social climate but also the daily lifestyles of a clear majority of the population, including large parts of the blue-collar worker families: In 1962 and 1963 the question about one s self-identification concerning lifestyle was not included in the survey. 7. See in this context Vogel s rich ethnographic description of the new middle class in urban Japan (Vogel 1965), in which he portrays Japanese society of the early 1960s as a still not affluent society.

8 12 David CHIAVACCI (1) The reinvention of the conservative establishment as champion of shared growth. (2) The establishment of the Japanese employment model. (3) The substantial amplification of mass consumerism and social upward mobility because of high growth. After the conflicts of , the new Prime Minister Ikeda Hayato not only set priorities back to a strong economy instead of rearming and a politically strong Japanese nation, but he was also able to establish a social consensus for this economic growth project (Chiavacci 2007: 37 39). Under Ikeda s leadership, the plan to double Japan s Gross National Product (GNP) that had been formulated by the former Kishi cabinet became known as the National Income Doubling Plan (Kokumin Shotoku Baizo Keikaku). This relabeling of the plan shows an important shift in the economic policies of the conservative establishment. Economic growth was no longer defined as a goal per se, but as an instrument to generate wealth for the whole population. Simultaneously and complementally, industrial relations were pacified through the institutionalization of the Japanese employment model (Gordon 1998: ). This Japanese employment model consisted of three pillars: implicit employment guarantee, seniority as important principle regarding pay and promotion and intensive cooperation between company labor unions and management; it was the basic compromise in industrial relations that became in the 1960s more and more dominant and led to a significant decrease in labor unrest and strikes. Through these two changes, the conservative establishment reinvented itself as champion of shared growth. Political priorities were not only set on promoting economic growth, but the conservative establishment guaranteed through its new rhetoric that the fruits of the economic efforts would be shared among the whole population. Moreover, the perception of lifestyle differences was fundamentally altered in those years. Thanks to high growth, mass consumerism penetrated fully Japanese society. Consumer goods, which had been confined to a small upper class, had become in a few years items that could be found in nearly all households. Social differences in lifestyle by social class did not disappear (Hara and Seiyama 2005), but the difference of a small wealthy minority owning a car and the majority without was of a completely different quality than the new difference of a small wealthy minority owning an expensive car and the large majority owning a much more modest model. Furthermore, because of high growth, social upward mobility strongly increased from the mid-1950s to the mid-1960s (Ishida 2001: 593). Many families experienced not only a rising lifestyle but better paid and more prestigious jobs became increasingly available for them and their offspring. A portrait in the US-American Time magazine of a young Japanese couple, both blue-collar workers at Matsushita, highlighted this new lifestyle and future aspirations even among blue-collar families: Evenings they watch Laramiè or the samurai dramas on their television set and fight off the winter chill by toasting their feet on an electric footwarmer. So well paid are their jobs...that they also own a refrigerator, transistor radio, vacuum cleaner, electric iron and washer. If the expectant Kumiko presents him with a son next month, Seiji even talks confidently of sending the boy to a university (Time 1962: 93). Class antagonism weakened, and class division became much more fluid during the 1960s. A general equality seemed to be realized by overall participation in the Japanese success story. Through the highly successful implementation of Ikeda s National Income Doubling Plan during the 1960s, the conservative establishment could redefine themselves as the champions of shared growth and a fair society beyond class antagonism. They gained, in contrast to their role in the mass society debate of the late 1950s, a completely new credibility as articulator of a general middle-class model. Through

9 From Class Struggle to General Middle-Class Society to Divided Society 13 the overall social upgrading of the Japanese society, the fuzzy metaphor of a general middle-class society gained more and more credibility in the lifeworlds. 3.3 Enhancement of the General Middle-Class Model During the 1970s, the character of the general middle-class model was enhanced in two aspects. First, it was consolidated and became much more concrete. Second, it changed its quality from a descriptive model to a prescriptive one Concretion and Consolidation During the 1970s, internationally comparative research attested that Japan is a model of equality regarding outcome as well as chances in comparison to advanced industrial economies of the West. Publications by the Organisation for Economic Co-Operation and Development (OECD) declared Japan to have a very equal income distribution per household (Sawyer 1976: 14, 17) and described Japan as a prime example for a credential society (gakureki shakai), in which educational success is a precondition for attaining high positions in society (Galtung 1971). Other comparative studies confirmed an equal income distribution in Japan per household (e.g. Boltho 1975: ). And leading foreign scholars of Japan praised its highly meritocratic school system as exemplary (Dore 1987: ; Reischauer 1977: ; Vogel 1979: ). These foreign studies were widely quoted in Japan and consolidated its societal model as a general middle-class society. In this consolidated view, Japan had not only a unique 90% middle-class consciousness but also could be explained by its exceptional degree of equality regarding chances and outcome in international comparison. The conservative establishment was not a passive spectator in the production of this enhanced general middle-class model of Japan. From the 1970s onward, the government actively fostered and reinforced the view of Japan as a uniquely equal and harmonious society (Taira 1993: ). Ministries and their research institutes published their own studies that supported a view of Japan as an exceptionally equal and fair society (e.g. EPA 1974). The model of Japan as a highly equal society complemented the discourse of Japan s cultural uniqueness (Nihonjinron) that was also promoted by the conservative establishment (cf. Befu 2001; Goodman 2005). One of the basic canons of this discourse was Japan s unique homogeneity as basis for its economic success. According to the general middle-class model, this meant not only a unique degree of ethnic homogeneity but also an outstanding degree of social homogeneity. Japanese social scientists were rather late to join the discourse of Japan as a general middle-class society. In 1977, Murakami Yasusuke published a small newspaper article that started a second debate about social stratification among social scientists. In his article, Murakami (1977) described Japan as a society dominated by a large middle class because of the self-identification of over 90% of the population with a middle position in the surveys of the CAO (cf. Murakami 1984). Other social scientists strongly questioned this view of Japan as a general middle-class society (Kishimoto 1978; Tominaga 1979). Still, even Murakami s critics started their argument from the premise of a distinctive general middle-class identification in Japan based on the surveys of the CAO. They argued that this general middle-class identification was due to a false class consciousness in Marxist terms or to a pronounced status inconsistency in Japan. Still, their acceptance of the general middle-class consciousness as distinct to Japan is an indicator of how dominant this idea had become at that time. However, empirical support for highly equal income distribution and for a highly open social stratification system in Japan in international comparison is not unequivocal. According to the well-known OECD study by Sawyer (1976) mentioned above, Japan is not only in the group of the countries with an equal income distribution around 1970 but the income share of the lowest 20% income group is in

10 14 David CHIAVACCI Table 1. Income Distribution per Household Around Sawyer (1976: 14, 16) Year Gini Income share of the lowest 20% (%) Ishizaki (1983: 72) Year Gini Income share of the lowest 20% (%) Sweden Sweden Norway Norway Australia Australia Japan UK UK Netherlands Netherlands Canada Canada Spain Spain USA USA West Germany West Germany Italy Italy Japan France France Source: Sawyer (1976: 14, 16) and Ishizaki (1983: 72). Japan significantly higher than in any other economy included in the study (Table 1). However, the study by Ishizaki (1983) shows opposing results. According to his calculations, Japan has very unequal income distribution in international comparison around 1970 and the income share of the lowest 20% is not especially high. 8 In research on social mobility, two indicators of mobility are employed: absolute mobility and relative mobility (Erikson and Goldthrope 1992: 54 59; Breen 2004: 3 4). Absolute mobility is the total of measured social mobility. However, absolute mobility of a society is strongly influenced by structural change and a number of other factors. For example, if a society experiences a fast industrialization, this leads to a transformation of its occupational structures and thus a high rate of absolute mobility. However, this cannot be regarded as an indicator that its social structures are more open in comparison to societies with lower absolute mobility. Therefore, many specialists regard relative mobility or social fluidity, which is based on differences in chances to attain a social status by social origin, as a better indicator for a comparative assessment of social mobility. According to empirical research on social mobility, postwar Japan has in international comparison with Western economies, a very high rate of absolute mobility (Erikson and Goldthrope 1992; Ishida 1993, 2001). This is not surprising in view of the high economic growth in Japan, which was clearly above the international average during the high-growth period as well as from the oil shock to the early 1990s. However, if relative mobility is considered, Japan s mobility rates are not above average (Erikson and Goldthrope 1992; Ishida 1993, 2001). According to a comparative study of Marshall Moreover, Sawyer (1976: 17 20) tries in a much less well-known second part of his study to increase the international comparability of the national data sets he uses. The calculations in this second part of his study do not report an especially equal income distribution for Japan in international comparison.

11 From Class Struggle to General Middle-Class Society to Divided Society 15 Table 2. Overall Relative Mobility. Men Women Estonia 0.49 USA 0.46 East Germany 0.35 East Germany 0.42 Czechoslovakia 0.30 Russia 0.26 Russia 0.21 Estonia 0.23 Bulgaria 0.19 Czechoslovakia 0.15 USA 0.05 UK 0.09 Poland 0.11 West Germany 0.18 Slovenia 0.15 Poland 0.19 UK 0.25 Bulgaria 0.21 West Germany 0.48 Japan 0.49 Japan 0.50 Slovenia 0.53 Source: Marshall (1997: 8). (1997), relative mobility in Japan is even very limited. Table 2 shows his findings concerning the overall relative mobility of men and women. Zero is the average rate of relative mobility among the countries included in the study. A negative value shows a higher rate of overall relative mobility and a positive value shows a lower rate of overall relative mobility. An explanation for this low relative mobility and strong rigidity of social stratification in Japan is the strong influence of social origin on educational attainment. In his extensive analysis of comparable data sets for the UK, the US and Japan, Ishida (1993: 66) concludes that the dependence of high school completion and college attendance on social origin in Japan appears to be larger than in the Unites States and Britain. Hence, the general middle-class model was strengthened and consolidated from the 1970s onward, but empirical support for its basic new assumptions was still very weak. Still, the consolidated general middle-class model could further strengthen its position and became fully dominant in the 1970s because it had a high empirical credibility in the lifeworlds. Relative social mobility is arguably the better indicator for internationally comparative research on the openness of social stratification systems, but in the lifeworlds only absolute social mobility is experienced. Even after the end of the high-growth era in 1973, absolute social mobility remained very high in Japan. The assumption of an open stratification system in Japan was confirmed in the general experience of social upward mobility in the lifeworlds. Moreover, the articulators of the model included not only the conservative establishment in Japan but also leading foreign researchers and renowned international organizations, which further increased the credibility of the model From Descriptive Model to Prescriptive Model From the 1970s onward, the general middle-class model changed its qualitative substance by being transformed from a descriptive model into a prescriptive model. Its core element as prescriptive model was the ideal of the sararıman (salaryman). An employment and a career in a large corporation (or as bureaucrat in the state administration) constituted the ideal life course embedded in the general middle-class model. Kariya Takehiko (1995: i) has summarized this ideal life course as follows:

12 16 David CHIAVACCI One attends a first-class juku, passes through first class middle and high schools and enters a top university. If one has achieved this, then one can enter a first-class company and have a happy life. Good education / good job / fulfilled life. In the Japanese postwar society this success story has been drummed into the people and spread as a life plan into every last corner of society. According to the general middle-class model, men could achieve this ideal life course regardless of his social origin, thanks to the meritocratic school system. Due to the strong segregation of gender roles in the postwar Japanese family model (Ochiai 2003), the ideal life course of women followed not the ideal sararıman life course for men. Instead, women as educational mothers (kyoiku mama) played the central part in the transmission of the ideal life course to the next generation. The general middle-class model together with the family model and the employment model constituted a prescriptive Japanese way of life (Chiavacci 2007: 41 45) according to which men study hard and become fully committed employees, and women marry a successful man and bear responsibility for transmitting this successful life path to the next generation. 9 The high resonance of this prescriptive general middle-class model was due to the fact that its ideal life course was for a significantly increasing share of the population actually achievable, thanks to a strong increase of new white-collar jobs. From the mid-1970s onward, the upper white-collar class strongly expanded (Ishida 2001: 593; 2002: 21). Among the male school graduates from 1946 to 1955, only 18.7% got jobs in the civil service or in a large corporation as a white-collar employee (Figure 2). But in the cohort of , this share had risen to over 30% and further expanded to over 40% in the cohort from 1986 to Furthermore, another fourth of these two cohorts could partly realize the ideal life course of the general middle-class model by finding work as white-collar employees in small- and medium-size companies. Even a part of the blue-collar workers could consider their job attainment as partial realization of the ideal life course, thanks to the whitecollarization (Koike 1988) of blue-collar workers as part of the Japanese employment model. Moreover, the general middle-class model started to exert a strong influence on social norms and the perception of society. Social egalitarianism emerged as a fundamental value and central pillar of the social contract of Japan and was shared by the population as well as by the elites. Several representative comparative surveys and studies of the 1980s and early 1990s show that the backing of egalitarian principles was significantly stronger and that ideological cleavages between the conservative establishment and left-wing opposition were smaller in Japan than in advanced industrial societies of the West (Verba et al. 1987: 76 82; Kluegel and Miyano 1995: 90 and 96; Wright 1997: ). The general middle-class model also resulted in a view of Japan as a credential society (gakureki shakai) par excellence. Graduation from a university was regarded as a prerequisite for a successful life. While only 22% of the respondents in 1951 and 38% in 1960 said that they would prefer to send a college-aged son to a university, this percentage had increased to 70% by 1973 (NHK 1975: 83). However, at the same time during the 1970s, the expansion of tertiary education temporarily halted, which made university education the bottleneck in the ideal life course. This led also to negative facets in the discourse of Japan as a credential society. Critics identified an overemphasis on educational credentials that resulted in excessive competition in the education system (e.g. Ogata 1976) With this turn, the discourse of the general middle-class model actually became contradictory in itself. On the one hand, it still proclaimed that everybody was equal, but on the other hand some were more equal than others (to paraphrase George Orwell). Still, this is not a specific phenomenon of Japan, but is embedded in every modern society that is democratic and a social market economy.

13 From Class Struggle to General Middle-Class Society to Divided Society % 8.7% 10.3% 71.0% % 15.3% 16.4% 61.0% % 18.5% 18.2% 53.2% % 21.5% 24.3% 43.4% % 31.2% 25.5% 33.3% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Civil Service White-collar employee in small and medium corporations White-collar employee in large corporations Self-employed, blue-collar worker, farmer Figure 2. Occupation of Male School Graduates, Source: Nakamura (2000: 50). The strong belief in the gakureki shakai as part of the general middle-class model shaped social research as well as public debate. The importance of social origin on educational attainments was neither in public debate nor was scientific research deemed an important subject (Kariya 1995: 55 57; Watanabe 1997: 60). Instead, the significance of educational credentials in the work career was a central topic in scientific analysis, and the strong and inhuman competition in school was a recurrent theme in public debate. From the 1970s onward, the general middle-class model was significantly concretized and gained a prescriptive character. Although empirical support for this model of Japan as an outstanding equal society regarding chances and outcome is very weak, it could retain and even strengthen its dominant position because of its positive resonance in the lifeworlds. The end of the period of economic high-growth led not to the end of the social upgrading of society. Not only did absolute social mobility remain very high, but the most important trend in the social stratification system was the increasing upward mobility into the upper white-collar class. This madethenewideallifecourseofthesararıman highly credible in the lifeworlds. Large segments of society could successfully be integrated into the model as the core element was actually achievable for them, at least partly. Moreover, the general middle-class model was a central element of the Japanese way of life as the mainstream of this period, which reinforced and stabilized its resonance in the lifeworlds. 3.4 Replacement of the General Middle-Class Model by the Divided Society Model From late 1990s onward, the general middle-class model came under strong pressure and has been replaced in recent years by a new societal model of inequality, which describes Japan as a divided society (kakusa shakai). At the origins of this new discourse regarding social inequality are two best

14 18 David CHIAVACCI sellers by Tachibanaki Toshiaki (1998) about income distribution and by Sato Toshiki (2000) on social mobility patterns. They were followed by a flood of studies and publications about social differentiation trends in current Japan. 10 In recent years, popular treaties of this topic with lurid titles have appeared, and major newspapers have published series about rising inequalities, the birth of a divided Japanese society and its consequences. Moreover, the divided society model has become a political issue and was a major factor in the success of the oppositional Democratic Party of Japan (DPJ) in the House of Councillors election in From Equality to Inequality? Has Japan been transformed from a highly equal society to a highly unequal society as the displacement of the general middle-class model by the divided society model suggests? The empirical studies by Tachibanaki (1998) and Sato (2000) as points of origin of the new model depict such a transformation of Japan and question the pillars of the existing general middle-class model. Tachibanaki (1998) challenges the assumed equality in outcome by showing a strong increase in income inequality per household from 1980 to 1992 and reaching the conclusion that Japan has the most unequal income distribution of all advanced industrial economies (Tachibanaki 1998: 6). Sato (2000) highlights new basic inequalities in chances. According to his study, relative chances to reach a position in the upper white-collar class had significantly decreased between 1985 and 1995 if one s father has not already occupied such a position (Sato 2000: 51 63). A social closure of the knowledge elite, as Sato called the upper white-collar class, has occurred. However, empirical studies on income distribution and social mobility do not support a fundamental transformation of Japan from a general middle-class society to a divided society. Regarding social mobility, some studies also report an increasing influence of social origin on study motivation and educational attainment (Kariya 2001: ; Kondo 2001). However, the large majority of studies show no significant change in social mobility patterns in Japan in recent years (Ishida 2001; 2002; Kanomata 2001; Hara and Seiyama 2005). Moreover, the study of Sato (2000) has been strongly criticized due to conceptual problems concerning his knowledge elite as well as the small data sample, on which his analysis is based (Seiyama 2000; Hashimoto 2003: ). Regarding income inequality, despite significant methodological progresses in measuring and internationally comparing income inequality since the 1970s, Japan s income distribution per household in international comparison remains an unsolved mystery. For example, depending on the study, Japan s position regarding income distribution among advanced industrial economies around 1990 varies widely (Table 3). According to the data published by the World Bank, Japan has one of the most equal income distributions of all advanced industrial economies. In contrast, the Luxembourg Income Study (LIS) reports for Japan one of the most unequal income distributions. The cause for these diametrically opposed results for Japan s income inequality lies in different national surveys. Economists mostly use two national surveys to measure income distribution per household in Japan: the Family Income and Expenditure Survey (FIES) or the Income Redistribution Survey (IRS). According to empirical studies that are based on the FIES (including World Bank), Japan has a very equal income distribution. However studies using the IRS (including LIS) find a very unequal income distribution in Japan. The findings of Tachibanaki (1998), who uses in his For overviews, see Bungei Shinsho Henshubu (2006) and Chuo Koron Henshubu (2001).