1 Introduction

Class mobility is an indicator of openness and equality of opportunity in society (Blossfeld & Shavit, 1992; Nennstiel, 2021, p. 1), therefore, there has been a large accumulation of class mobility studies, especially on social stratification. There have been discussions on the consequences of class mobility for individuals since the 1950s (Blau, 1956; Sorokin, 1959). This is because class mobility is not only a transition from an origin class to a new destination class but is accompanied by social, economic, and cultural changes.

The literature has indicated that class mobility affects social and political behaviors, values, and subjective well-being (SWB). Blau (1956) suggested that all new members of a social class have difficulty acculturating to its values, lifestyles, and behaviors, all of which differ from those of non-mobile members, and that the process also causes social insecurity (Blau, 1956, pp. 290–91). Sorokin (1959) assumed that class mobility entails mental strain and illnesses which cause psychological distress.

Although class mobility research has been thoroughly documented, empirical evidence of the consequences of class mobility for mobile members has been scarce (Marshall & Firth, 1999; Nikolaev & Burns, 2014). While Blau and Sorokin’s works appeared in the 1950s, only a few empirical studies have been conducted since then (Alston & Knapp, 1974; Mitchell, 1972). Since the 2000s, empirical research conducted on the relationship between class mobility and SWB (i.e., life satisfaction, happiness, self-rated health) has started to accumulate. However, empirical evidence has been mixed. Some studies found no significant relationship between class mobility and SWB (Dhoore et al., 2019; Houle, 2011; Marshall & Firth, 1999). Whereas several studies reported significant effects of class mobility, the observed effects were inconsistent. Some indicated a positive (negative) impact of upward (downward) mobility on SWB (Nikolaeve & Burns, 2014; Zhao et al., 2017; Chan, 2018), whereas others indicated a negative (positive) effect of upward (downward) mobility (Graham & Pettinato, 2006; Hadjar & Samuel, 2015; Zang & Dirk de Graaf, 2016). The previous research remains inconclusive on the relationship between class mobility and SWB.

Although the literature has indicated that the consequences of class mobility can differ based on the macro-social context, research focusing on macro-social factors has been scarce. According to Germani (1966) and Goldthorpe (1987), high mobile rates may ease the detrimental consequences of class mobility, and those who shifted to a new class in society in large numbers experienced fewer issues while adapting to the destination class. As mobile individuals rely on several other members who have had similar experiences of mobility into their destination classes, they generally find it easy to integrate into their new class without experiencing social isolation. These studies show that the relationship between class mobility and SWB depends on mobility rates.

A few studies have considered the macro-social context to examine the relationship between class mobility and SWB (Dhoore et al., 2019; Hadjar & Samuel, 2015; Marshall & Firth, 1999). Marshall and Firth (1999) and Dhoore et al. (2019) concluded that there is no evidence that class mobility and SWB differ based macro-social context, whereas Hadjar and Samuel (2015) concluded that the effect of upward class mobility on SWB differed between countries. As most of these studies have focused on Western society, it is difficult to generalize the results concerning the impact of macro-social factors on the relationship between class mobility and SWB. Several class mobility studies have focused on the relationship between class mobility and the macro-social aspects of industrialization (Erikson & Goldthorpe, 1987, 1992; Erikson et al., 1982; Featherman et al., 1975; Hout & Jackson, 1986; Ishida, 1993; Ishida et al., 1991; Lipset & Bendix, 1959). However, little is known about the consequences of class mobility and how they are affected by macro-social factors.

This study considers the macro-social context and focuses on the relationship between class mobility and SWB. It examines intergenerational social class mobility, which indicates mobility from the father’s social class to the respondent’s present social class, and life satisfaction, which is one of the representative measurements of SWB. It investigates whether the association between social class mobility and SWB depends on socioeconomic factors. It employs the Diagonal Reference Model (DRM) created by Sobel (1981) to simultaneously estimate the origin, destination class, and effects of mobility.

2 Japan as a Case Study

This study focuses on Japanese society, which was the earliest Asian nation to industrialize. Understanding the mechanism of SWB in Japan, which has not been examined thus far in the literature, will facilitate a broad understanding and provide a valuable perspective on the relationship between class mobility and SWB. First, Japan has experienced remarkable economic changes over time. After World War II, it achieved a significant economic recovery and entered a phase of rapid economic growth from the mid-1950s to the early 1970s. Its real GDP and GDP growth rates increased rapidly during this period (Fig. 1). After the period of high economic growth, while the rate of growth slowed, it remained stable from the 1970s to the 1980s. When the bubble economy collapsed in the early 1990s, a prolonged economic recession began.

Fig. 1
figure 1

Socioeconomic changes in Japan. Note) Real GDP indicators are from: MLIT (2012), Cabinet office (2021), High school/ college admission rate are from: e-Stat (2021), Ageing rate: Statistics Bureau of Japan (2020), Fertility rate: MHLW (2020), Blue (the sum of skilled manual, non-skilled manual), White (the sum of professional-managerial, routine, non-manual) collar, and Farming rates: Ishida (2018)

Full size image

Along with economic changes, Japan’s educational and occupational structures also changed significantly. As Fig. 1 shows, along with the expansion of compulsory education since the 1950s, the average educational level increased significantly. With progressive industrialization, the agricultural sector shrank dramatically, while both white and blue-collar occupations began to expand.

Demographic characteristics have also changed drastically since the postwar era. The size of the elderly population increased from 5.3% in 1955 to 28.9% in 2020 (Statistics Bureau of Japan, 2020), which is the highest in the world. The total fertility rate declined from 2.37 in 1955 to 1.36 in 2019 (MHLW, 2020).

These significant socioeconomic changes, which have occurred only within 70 years, are inevitably related to social class mobility. The total mobility rates increased from the 1950s to the 1970s (Tominaga, 1979, 1992), wherein many members of society left the agricultural sector and shifted to the industrial sector. After the high economic growth period, the social structure gained stability. However, since the recession (from the 1990s), the rates of upward mobility have decreased while downward mobility has increased (Ishida, 2018, p. 49).

Japan has experienced significant socioeconomic change and with these changes, both social class and patterns of mobility have also changed. Therefore, the nation is most suited for an examination of how macro-social factors affect the relationship between social class mobility and SWB. As only a few studies have examined this relationship, little is known about how class mobility affects SWB and whether it has changed with socioeconomic changes in Japan. Kanai (2018), one of the few to examine this relationship, found that intergenerational downward class mobility began to affect SWB in the 2010s. This implies that socioeconomic changes affect the relationship between SWB and class mobility.

An additional advantage is that Japan has highly consistent datasets (Social Stratification and Social Mobility Survey) in place from 1955 to the present. These data allow us to track the impact of socioeconomic changes over time. Examining the relationship in one country can avoid measurement errors caused by cross-national analysis (Kang et al., 2020). Using Japan as a case study, will provide valuable insights for both Western countries in general and other Asian countries that have experienced similar processes of industrialization.

3 Literature Review and Hypotheses

This section reviews the theoretical and empirical research on class mobility and SWB and formulates hypotheses accordingly. Blau (1956) initially suggested the acculturation hypothesis, which assumes that mobile members’ behaviors, values, and political behaviors are between the origin class and the destination class. Mobile members are not well integrated into the new class lifestyle. They do not have intensive contact with the new class; therefore, they do not have substantial opportunities for complete acculturation to the new class, either (Blau, 1956, p. 3). Mobile members are affected by both the origin and destination classes, and their behaviors are between the two. The acculturation hypothesis does not postulate an additional independent mobility effect.

Some empirical studies have reported results consistent with Blau’s acculturation hypothesis (Dhoore et al., 2019; Houle, 2011; Houle & Martin, 2011; Schuck & Steiber, 2018). They have suggested that the SWB of mobile members lies between the non-mobile members of the origin and destination classes. None of them found any additional mobility effects. Therefore, considering Blau’s acculturation hypothesis and following prior empirical evidence, we can formulate Hypothesis 1 as follows:

H1. The SWB of mobile members is affected by both the origin and destination classes. As class mobile members never fully acculturate to the destination class, the SWB of mobile members lies between the non-mobile members of the origin and the destination classes (the acculturation hypothesis).

In contrast, Sorokin (1959) assumed that mobility accompanies the independent negative impact on the SWB of mobile members. This is called the dissociative hypothesis. Mobility accompanied massive economic and cultural changes. Mobile members experienced different environments and had to adapt to new settings, both cultural and economic. Sorokin found that these changes caused permanent mental strain and stress for mobile members. Therefore, mobility itself, regardless of the direction, has adverse effects on the SWB of a mobile member of society.

Studies have been conducted to empirically confirm Sorokin’s dissociative hypothesis. However, there has been little support for it. Most studies have not found empirical evidence of the dissociative hypothesis (Chan, 2018; Dhoore et al., 2019; Houle, 2011; Houle & Martin, 2011; Zhao et al., 2017). Only a few found partially dissociative effects (Ellis & Lane, 1967; Hadjar & Samuel, 2015). Despite the scarce empirical support, Sorokin’s hypothesis is one of the most critical theoretical approaches toward considering the relationship between SWB and social class mobility. What insights does Sorokin’s hypothesis provide on the relationship between SWB and hierarchical mobility and its changes in the relationship in Japan? We formulate Hypotheses 2 and 2a as follows:

H2. SWB has an additional mobility effect.

H2a. All mobilities have adverse effects on SWB because mobility itself accompanies changes in the cultural or economic background, which cause strain and stress regardless of the direction of mobility (the dissociative hypothesis).

Furthermore, Duesenberry (1949)’s relative income hypothesis has implications when considering the impact of social class mobility on the SWB of mobile members. According to Duesenberry, utility is not determined by the absolute level of an individual’s expenditure but rather by a function of the ratio of current expenditure to others (Duesenberry, 1949; Easterlin, 1974). The current utility is determined by a “comparison” to the reference group. The reference point can be distinguished into an internal reference point such as the past or expected future income and an external reference point such as family, other workers, region, neighborhood, or country (Clark et al., 2008, p. 100). Duesenberry’s relative income hypothesis has been applied in SWB research in the context of “social comparison” (Easterlin, 1974, 1995; McBride, 2001; Clark et al., 2008; Clark & D’Angelo, 2010).

By applying Dusenberry’s relative income hypothesis to the relationship between social class mobility and SWB, social class mobility can bring forth two different consequences on SWB. First, if a comparison is made with a past internal reference point such as the origin class, moving upward can affect a mobile individual’s SWB positively, whereas moving downward can affect their SWB negatively. Individuals determine the SWB based on changes in their relative social status.

Some empirical studies have shown results that are consistent with this mechanism. Nikoleave and Burns (2014) and Chan (2018) have reported that upward mobility affects SWB positively. Zhao et al. (2017) and Kanai (2018) reported that downward inter/intra-generational mobility harms SWB.

If an individual determines the SWB based on internal reference points, we can formulate Hypothesis 2b as shown below. We refer to this as “the relative status hypothesis” because it assumes that changes in relative status affect SWB.

H2b. SWB is determined by a comparison of the past and present statuses. Therefore, whereas upward mobility has a positive impact on SWB, downward mobility has a negative impact (the relative status hypothesis).

If mobile individuals determine SWB through an external reference group, that is, non-mobile members of their destination class, upward mobility can have a negative effect because of their less advantageous origin class, and downward mobility can have a positive effect as their relatively beneficial origin class. Graham and Pettinato (2006) identified upwardly mobile members of middle-income group in Peru have relatively low SWB, and they called them “frustrated achievers.” Zang and Dirk de Graaf (2016) found that intra-generational downward mobility positively affected SWB and called those presenting this condition “satisfied losers.” Graham and Pettinato (2006) and Zang and Dirk de Graaf (2016) suggested that social comparison can affect SWB through a comparison with non-mobile members of the destination class.

If SWB is determined based on an external reference point, Hypothesis 2c can be postulated as shown below. In this study, we refer to this as “the inertia hypothesis” because it means that the economic and cultural background of the origin class will continue to affect mobile members.

H2c. The comparison with non-mobile members in the destination group determines SWB. Whereas upwardly mobile members harm SWB because of their less beneficial origin class, downwardly mobile ones have a higher SWB because of their beneficial origin class (the inertia hypothesis).

When we examine the SWB of members of society, we must consider a macro-social context. Changes in socioeconomic structure can affect people’s well-being. Research suggests that high social mobility rates diminish the detrimental consequences of social mobility. As many people experienced social mobility during the period of dramatic change in socioeconomic structure, mobile members did not have severe maladjustment problems because there were many of them in their destination class. In contrast, when few socioeconomic changes occur and the social structure remains entrenched, the problem of adaptation to a new class may become prominent (Dhoore et al., 2019; Germani, 1966; Goldthorpe, 1987; Kessin, 1971).

Several studies that have considered the macro-social context have suggested that macro-social factors have an independent impact on SWB (Alesina et al., 2004; Senik, 2006; Fischer, 2009). Living in a socially mobile society (Fischer, 2009) or a society that is perceived to have high social mobility (which may not necessarily be true) (Alesina et al., 2004; Senik, 2006; Fischer, 2009) has been found to have a positive impact on an individual’s SWB. Kanai (2018) found that intergenerational downward mobility harms life satisfaction among those who participated in the labor market after the 1990s.Footnote 1 Given this, we can assume that social mobility effects on SWB can differ according to the socioeconomic experience of members of society.

How will the macro-social context change the relationship between SWB and intergenerational mobility in Japan? We divide the time period into three phases of economic development: high economic growth, stable economic growth, recession eras, and examining the relationship between SWB and intergenerational social class mobility. In Japan, both socioeconomic structure and intergenerational mobility are closely related to the stage of economic development. Intergenerational social class mobility increased especially during the period of high economic growth (Tominaga, 1979, 1992). It then increased gradually in the 1980s and 1990s (Hashimoto, 1999), which was the period of stable growth. The number of informal jobs increased significantly during the recession (indicating that more young people moved into the lower classes) and barriers to interclass mobility increased (Hashimoto, 1999, 2018; Sato, 2000). Considering the prior literature and the Japanese context, we present Hypothesis 3 as follows:

H3. The effect of social class mobility on SWB differs based on socioeconomic changes and the extent of social class mobility. The detrimental impact of social class mobility was weak during the period of high economic growth because many people experienced mobility. However, the impact became vital during the recession because mobility decreased.

4 Data and Variables

4.1 Data

This empirical study is based on the National Survey of Social Stratification and Social Mobility (SSM) datasets of 1985, 1995, 2005, and 2015. This survey has been conducted by Japanese sociologists once every 10 years since 1955, which is one of the representative surveys in Japan. Sampling is conducted by each survey year based on the Basic Resident Register. It includes information on the socioeconomic status of both the respondents and their fathers. The survey has adopted many consistent questions throughout, which has made tracing change across time possible. In this paper, we integrated all datasets from 1985Footnote 2 to 2015 and used data for males and females aged 30 to 54 years who were born after 1936. We ensured that our research subjects were aged under 55 years to exclude people who re-entered the labor market after retirement. We also eliminated individuals who did not have any occupation at the time of study. After these criteria were applied, 9055 cases were found valid.

4.2 Variables

4.2.1 Dependent Variables

The dependent variable is SWB, which refers to people’s cognitive and affective evaluations of their lives (Diener et al., 2009). SWB has been included in questionnaires for decades since Bradburn (1969), Cantril (1965), and Likert (1932) (Ferrer-i-Carbonell, 2005, p. 1002). Many empirical studies have used self-reported measures of well-being such as life satisfaction and happiness (Dhoore et al., 2019; Hadjar & Samuel, 2015; Marshall & Firth, 1999; Nikolaev & Burns, 2014; Zhao et al., 2017).

Self-rated well-being may not be a perfect measure because it is based on the respondent’s self-assessment. Regardless of the limitations, previous studies have shown that it is a reliable and consistent measure that contains important information (Costa & McCrae, 1988; Sandvik, 1993; Fernádez-Dols & Puiz-Belda, 1995; in details, see Nikolaev & Burns, 2014, p. 85). This study uses self-rated life satisfaction, which is the cognitive component of SWB based on evaluating the past, present, and future conditions (Campbell, 1981; Diener et al., 2005; Hadjar & Samuel, 2015, p.49). Considering that estimated happiness and life satisfaction equations have similar structures (Blanchflower & Oswald, 2004), using life satisfaction as a representative measurement of SWB is reasonable. The question on life satisfaction was: “Overall, are you satisfied with your life?” The reply options on a Likert scale were: 1. Satisfied, 2. Somewhat satisfied, 3. Neither, 4. Somewhat dissatisfied, 5. Dissatisfied. We used these variables as continuous and reversed the options to interpret the results easily. Therefore, the options are: 1. Dissatisfied–5. Satisfied.

4.2.2 Independent Variables

The independent variables in this study include the origin and destination classes and mobility variables. The EGP class scheme comprising the following six categories (Erikson & Goldthorpe, 1992) was used: (1) service class, (2) routine clerical/sales, (3) petty bourgeoisie, (4) skilled manual, (5) semi-unskilled manual, and (6) farmers/farmworkers. The origin class refers to the main occupation of the respondent’s father. The destination class refers to the respondent’s occupation at the time of study.

We applied the “dominance” approach (Breen, 2004) for the married double-worker to determine a single position for a household. This approach uses two criteria: a full-time worker dominates a part-time worker and if both occupational positions are the same, the service class and the petty bourgeoisie dominate other classes. The six-category version of the EGP class scheme using SSM data was coded according to Takenoshita et al. (2008).

The EGP class scheme does not assume a class hierarchy. To capture the directions of mobility, we drew from Houle (2011) and used three mobility variables, namely upward, downward, and horizontal mobility. Upward mobility refers to mobility from another class to the service class or petty bourgeoisie, downward mobility refers to mobility from the service class or the petty bourgeoisie to another class. Horizontal mobility refers to all other forms of mobility. Accordingly, there were 2550 non-mobile (28.2%); 1428 downwardly mobile (15.8%); 2277 upwardly mobile (25.1%); and 2800 horizontally mobile (30.9%) members.

4.2.3 Covariates

We included covariates that are related to dependent variables. To examine the changes in SWB, we considered the influence of the Age Period Cohort (APC) effect (Blanchflower & Oswald, 2008; Taromaru, 2016) on life satisfaction. Therefore, this study controlled the APC effect while using integrated data.

To control the age effect, age and squared age variables were used. To avoid multicollinearity, we used centered and centered squared age variables. To capture the cohort effect, we used three cohorts: that consider the economic circumstances when an individual gets their first job (20 years old), for persons born between 1936 and 1953 (high economic growth period), 1954 and 1970 (stable economic growth period), and 1971 and 1985 (recession period). To capture the period effect, we used the unemployment rate as a proxy variable. When we estimated the APC effect simultaneously, linear dependence occurred. There is a relationship among these variables, which is known as the “Identification problem” (Firebaugh, 1997). Besides the APC effect, gender and marital status were controlled.

5 Method

We adopted Sobel’s (1981) DRM. Prior research has indicated that it is not enough to determine the effect of class mobility using the traditional regression model while estimating mobility effects. As class mobility is measured by the difference between the origin and destination classes, a multicollinearity problem occurs. We cannot estimate the effect of origin and destination classes, and class mobility simultaneously (Hendrickx et al., 1993; Houle, 2011; Zhao et al., 2017, p.58). To solve this problem, Sobel (1981) suggested the DRM, which is based on Blau’s acculturation hypothesis. It postulates that an individual’s current behaviors, perceptions, and values are affected by both their origin and destination classes (Blau, 1956; Sobel, 1981, p. 895). Therefore, in the DRM, the behavior of mobile members is determined by the weight of two non-mobile members: the origin and destination classes.

Figure 2 presents a class mobility table. The orange cells represent non-mobile members, whereas the gray ones (diagonal ones) represent mobile members. The parameters of mobile members comprised the weights of non-mobile members of both the origin and destination classes. The parameter of mobile members was constrained to add up to 1. Therefore, the base model of the DRM was as follows.

$$y_{ijk} = p*u_{ii} + \left( {1 - p} \right)*u_{jj} + \varepsilon_{ijk}$$
(1)

\(y_{ijk}\) means outcomes \(k\) of an individual whose origin class is i and whose destination class is \(j\). \(p\) is the weight of the non-mobile origin class member and \(1 - p\) represents the weight of the non-mobile destination class member.\(u_{ii}\) indicates the population means of \(k\) of the non-mobile members of the origin class, \(u_{jj}\) indicates population means of \(k\) of the members of the destination class. \(\varepsilon_{ijk}\) is a stochastic term with expectation 0. If \(p\) is larger than 0.5, the effect of the origin class is bigger than that of the destination class, and if \(p\) is less than 0.5, the effect of the destination class is bigger than that of the origin class. We can also test whether there is an additional independent mobility effect over and above the origin/destination effects. We can add the mobility variables to estimation (1). The estimation is as shown below:

Fig. 2
figure 2

Class mobility table

Full size image
$$y_{ijk = } p*u_{ii} + \left( {1 - p} \right)*u_{jj} + \gamma 1down_{ij} + \gamma 2up_{ij} + \varepsilon_{ijk}$$
(2)

We used estimations (1) and (2) with additional covariates as the model. In Model 1, we measured the impact of the origin and destination classes on mobile members after controlling for covariates (APC effect, gender, and marital status) (estimation (3)). In Model 2, in addition to Model 1, we added mobility variables (upward, downward, and horizontal mobility) to test whether there is an independent effect of the direction of mobility over and above the acculturation influence (estimation (4)).

$$y_{ijk} = p*u_{ii} + \left( {1 - p} \right)*u_{jj} + \sum \beta cov_{ijk} + \varepsilon_{ijk}$$
(3)
$$y_{ijk} = p*u_{ii} + \left( {1 - p} \right)*u_{jj} + \gamma 1down_{ij} + \gamma 2up_{ij} + \gamma 3horizontal_{ij} + \sum \beta cov_{ijk} + \varepsilon_{ijk}$$
(4)

We used constrained nonlinear regression commands in SPSS to estimate the models.

6 Results

6.1 Social Class Mobility of Postwar Japan

Before clarifying the impact of social class mobility on SWB, we confirmed social class mobility among the cohorts. Table 1 displays the absolute amounts of non-mobility/mobility among cohorts. Table 1 shows that the percentages of mobility in the early two cohorts are almost the same. The result was quite different in the youngest cohort: whereas downward mobility increased, upward mobility decreased. The youngest cohort, whose members started their careers in the recession, experienced greater downward and lower upward mobility when compared to any of the previous cohorts. Table 1 presents the results of absolute social class mobility. Research has shown that relative social class mobility was remarkably stable in the postwar period in Japan (Ishida, 2018). Whereas relative mobility has not changed significantly, the youngest generation has had a greater chance to experience non-mobility and downward mobility than the former generations.

Table 1 Social class mobility among cohorts
Full size table

6.2 Social Class Mobility and SWB

Table 2 presents the average score of life satisfaction by class mobility table. The bold-highlighted cells placed diagonally represent immobility. All other cells represent mobility. Each cell represents an average score. The numbers in “n” indicate the number of respondents in each cell.

Table 2 SWB by social class mobility
Full size table

In Table 2, it can be seen that there is no obvious pattern of increase/decrease in SWB with class mobility. However, members of the service class tended to have relatively high scores for life satisfaction, whereas those of the semi-unskilled class tended to have lower SWB, regardless of their origin class. The SWB of the mobile members from both classes was similar to that of non-mobile members of their destination classes.

In the next section, we used DRM analysis and examined the components of the SWB of mobile members.

6.3 Multivariate Results

6.3.1 The Overall Effects of Social Class Mobility on SWB

Tables 3 and 4 present the results of the overall effects of social class mobility on SWB. As Table 4 shows, the average scores for non-mobile members differ across the classes.

Table 3 The overall effects of social class mobility (descriptive)
Full size table
Table 4 The overall effects of social class mobility (diagonal reference model)
Full size table

The average scores of life satisfaction for non-mobile members are ranked in the following order from high to low: service class (u11), routine clerical/sales (u22), petty bourgeoisie (u33), skilled manual (u44), farmers/farm workers (u66), and semi-unskilled class (u55). Whereas the service class had the highest score (3.541), the semi-unskilled class had the lowest (3.135). Aside from this, there was no consistent relationship between non-mobility and SWB.

From the covariates and the APC effect, it is clear that age and age squared were significant in life satisfaction (0.006, 0.001, respectively). Age had a curvilinear impact on life satisfaction (Blanchflower & Oswald, 2008; Kobayashi, 2016). The female dummy had a positive effect on SWB (0.198), which means that females had higher life satisfaction than did males if other variables were controlled. The marriage dummy was positively related to SWB (0.297). Married people had higher levels of life satisfaction than those who were single, divorced, or widowed. Cohort effects were significant. Younger cohorts had higher SWB than those born between 1936 and 1953. Life satisfaction for the youngest cohort was much higher than that for those born between 1936 and 1953 and who got their first jobs in the high economic growth era. Although they did not have a beneficial economic background and experienced greater downward mobility, they tended to be satisfied with life overall. Unemployment rates, which serve as a proxy variable for a period, were not significantly related to SWB. These results show that age and cohort factors, rather than a period factor, significantly impacted SWB.

Focusing on the origin, the destination weights in Model 1, the weight of the destination class is much larger (0.745) than that of the origin class (0.255). The destination class weight was significantly larger than 0.5, which means that the SWB of mobile individuals was mostly shaped by their destination and not origin class. However, as Model 2 shows, there was no additional mobility effect on SWB.

These results are consistent with Marshall and Firth (1999), Houle (2011), Hadjar and Samuel (2015), and Dhoore et al. (2019), all of whom reported no additional mobility effects on SWB. As Blau (1956) assumed, the SWB of mobile members in Japan was in the acculturation process and influenced by both the origin and the destination classes. The SWB of mobile members was rather similar to that of non-mobile members in the destination class. However, there was no empirical evidence of additional mobility effects over and above the origin/destination class. Therefore, H1 (the acculturation hypothesis) was supported, and H2 was rejected.

6.3.2 Cohort Analysis I: Impacts of Social Class Mobility on SWB Among Cohorts

Society includes different generations. Although all generations live in the present and in the current socioeconomic structure, each one has a different socioeconomic background because each has lived in a different era.

After World War II, Japan underwent drastic socioeconomic changes. During the period of rapid economic growth, Japan’s socioeconomic structure changed significantly. The peasant class disintegrated rapidly, and white-collar employment boomed (Hashimoto, 2013; Ishida, 2018). During the high economic growth period, most people were upwardly mobile, regardless of their absolute/relative mobility, and their wages and living standards also increased. Once the high economic growth period ended, the socioeconomic structure stabilized. After the bubble economy collapsed, Japan faced a prolonged recession. The employment market began to shrink and the number of non-regular workers increased, especially among the younger generation (Hashimoto, 2018).

Each generation in Japan has had a considerably different socioeconomic background and social class mobility experience. The youngest cohort had mobility experiences that differed from those of the older cohorts in the preceding section, and cohort experienced greater downward mobility. These different socioeconomic situations and experiences of mobility may have different consequences for SWB. Therefore, additional research is necessary to clarify whether mobility effects on SWB have changed with changes in the economic environment. This section conducts a DRM analysis for each cohort.

As shown in Model 1, the effects of the origin and destination classes and mobility on life satisfaction differed across cohorts. For the first cohort, the destination class weight was 1.0 and the origin class weight was 0, which means that the SWB of the male mobile members of these cohorts were not significantly different from that of the non-mobile members of the destination class. Although the destination class weight decreased in the second cohort, it dominated over the origin class weight. The life satisfaction of the mobile members of these two cohorts were mostly similar to that of the members of the non-mobile destination class.

However, the results for the youngest cohort, namely those who were born between 1971 and 1985 were very different from those of the older cohorts. The weights of the destination and origin classes decreased and increased considerably, respectively. The origin class weight (0.482) became more significant than that of the destination class (0.518). The life satisfaction of the youngest cohorts was formed by both their destination and origin classes. Figure 3 shows the trend in the origin/destination class weights. Whereas the impact of the origin class increased, that of the destination class decreased across cohorts. The mobile members, those who got their first job in the recession era, were more affected by their origin class (Table 5).

Fig. 3
figure 3

Trends in the origin and destination class weights (both males and females)

Full size image
Table 5 Cohort analysis
Full size table

Second, from Model 2, which validates the additional mobility effects, we found a significant mobility effect in the youngest cohort alone. Horizontal mobility harmed life satisfaction, whereas downward and upward mobility did not affect SWB for this cohort. However, aside from this, there were no significant mobility effects in any of the other cohorts.

Each cohort was affected differently by social class mobility. The high economic growth cohort was affected only by the destination class. We could not find any effect of origin class or mobility. Therefore, all hypotheses were rejected in the first cohort. In contrast, those who entered the labor market in the period of stable economic growth began to be affected by the origin and destination classes, although the origin class effect was rather small. Therefore, H1 was supported for the stable economic growth cohort. For the low growth cohort, the results proved H2, which postulated an additional mobility effect. However, the effect we found was not in perfect alignment with H2a, H2b, and H2c. We found that only horizontal mobility had an independent effect on SWB.

Horizontal mobility in this study includes movement from non-manual (manual) to manual (non-manual) occupations. Although we assumed that these mobilities belonged to the same category, there are considerable differences between non-manual and manual occupations in Japan. Horizontal mobility accompanies considerable cultural and occupational changes, without any significant socioeconomic rewards or disadvantages. It may cause some apparent confusion or stress, which may, in turn, negatively affect SWB for the young generation, who consider their origin class rather significant. Our results show that both the origin class and mobility began to affect the SWB of the younger generation.

The cohort analysis confirmed that social class mobility was considerably different for each cohort. Whereas the destination class effects tended to be dominant in the high economic growth era, the origin class effect tended to be significant in the recession era. Therefore, our results strongly supported H3.

6.3.3 Cohort Analysis II: Gender

We have confirmed the impact of social class mobility on SWB for the younger cohort, which had a more significant origin class effect. Social class mobility effects in the youngest cohort were significantly different from those of the former cohorts. However, this trend and mechanism may differ by gender. The career paths of males and females differ considerably in Japan. Many females are left out of the labor market after marriage and/or childbirth, and some of them re-enter the labor market after the infant care phase (Yoshida, 2004). In many cases, they enter the labor market as part-time workers to supplement their household incomes (Hashimoto, 2018). This section conducts a gender analysis and compares the mobility effect between males and females.

Tables 6 and 7 present the results for males and females, respectively. Model 1 in Table 6 shows that the results of the former two cohorts are significantly similar, and that the destination and origin class weights are 1.0 and 0, respectively. In contrast, as seen in the previous section, the result of the youngest cohort for males is considerably different from both these cohorts. The origin and destination class weights of the youngest cohort are 0.563 and 0.437, respectively. The origin class weight reversed the destination weight. Mobile members in the youngest cohort were more affected by their origin class. We were not able to find any mobility effect over and above these in any cohort for males. All hypotheses were rejected for the early two cohorts of males, whereas H1 was supported for the youngest cohort of males.

Table 6 Cohort analysis (Males)
Full size table
Table 7 Cohort analysis (Females)
Full size table

As can be seen in Table 7, the trend in the impact of social class mobility for females was quite different from that for males. The averages of origin class weights were considerably more extensive and the destination weight was smaller than that of males. In the previous section, we confirmed that the destination class weight for males born between 1936 and 1953 was 1.0. However, the destination class weight in the earliest cohort for females was much smaller. The trend in the impact of social class mobility was confirmed for the female cohorts; the younger cohort had considerable origin class weight. The origin class impact already expanded in the second cohort; that is, for those who got their jobs during the period of stable economic growth. It contrasts with the results found in the previous sections for males.

From Model 2 in Table 7, we confirmed additional mobility effects for the youngest cohort alone. Horizontal mobility had adverse effects on the SWB of the youngest female cohort. When other covariates and the origin/destination classes were controlled, horizontal mobility harmed SWB. We can thus assume that the negative horizontal mobility effect in Table 5 (analysis for both males and females) was mainly caused by the female cohort. H1 was supported for the first and second female cohorts. H2 was supported for the youngest cohort. However, H2a, H2b, and H3 did not explain this result fully.

The gender analysis presents three main findings. First, the trends in social class mobility impacted both males and females. The higher origin class influenced the younger cohorts. Second, the origin class impact was much larger among females of all cohorts than among males, and the origin class began to expand its affluence for females in the early period (see Fig. 4). Third, there were additional mobility effects in the youngest cohort of females alone. Horizontal mobility harmed SWB for the youngest females. The gender analysis also showed that the effect of horizontal mobility on the younger cohort was mainly attributed to females.

Fig. 4
figure 4

Trends in the origin and destination class weights (left: males, right: females)

Full size image

7 Discussion

Overall, the SWB of mobile members in Japan is under acculturation influence, just as Blau (1956) had assumed. The SWB of mobile members is shaped by their origin and destination classes. However, we did not find any evidence that all mobility has additional negative effects regardless of direction, as Sorokin (1959) suggested. These overall effects vary by cohort and gender. The relationship between intergenerational mobility and SWB is more variable and complex than existing studies have assumed. Whereas many existing studies have not considered the variations in socioeconomic factors and the impact of gender, this study clearly shows that the effect of intergenerational mobility on SWB varies based on both socioeconomic context and gender.

We did not find any evidence that mobile members, especially male ones who entered the labor market during the period of high economic growth had adjustment problems as a result of their mobility. They were found to have SWB that was not different from that of their counterparts from the non-mobile destination class. The results are consistent with Germani (1966) and Goldthorpe (1987), who asserted that high mobility rates may ease the detrimental consequences of class mobility.

In contrast, the impact of the origin class increased significantly for the youngest cohort. Considering both the literature and the Japanese context, this may be interpreted as Sudo (2009) did. Sudo (2009) focused on non-mobile members who had the same occupation as their fathers at the time of his study and found that upper white-collar succession positively affected subjective social status. He indicated that in the stable economic growth era, occupational values also gained stability. He also noted that people could assess the value of both their own current and their fathers’ occupations. Therefore, succession to an advantageous occupational position contributes toward having an “upper” status.

From this study, we can assume that a similar mechanism unfolded in the context of the SWB of the younger generation. During the recession, Japan’s economic growth rate was meager, and occupational value did not change significantly. However, the income gap increased, and the labor market also changed. The proportion of non-regular workers increased dramatically, and most non-regular occupations began to be filled by the young generation (Hashimoto, 2018). The most recent cohort was at a disadvantage when compared to the former cohorts. We can assume that the value of the origin class gained clarity in the present period and became a crucial factor in determining the SWB. In this socioeconomic situation, the financial, cultural, and family background of the origin class began to influence the SWB of young mobile members. In contemporary Japan, the young generation began to maintain its origin class legacy.

Our study indicated that the origin class and mobility were more critical factors for females, especially for the youngest among them. Japanese females followed more complex career paths than males. Marriage and childbirth are critical variables in the career path for females. Many females experience career breaks as they leave the labor market after marriage and/or childbirth and re-enter it as part-time workers once they have settled down to raise their children. Prior studies thus identified the status of their husbands or households as critical to determining the subjective status of married females, rather than their own individual incomes or statuses (Akagawa, 2000; Aramaki, 2019). The recent economic downturn and worsening labor market conditions have directly affected females whose careers were precarious from the beginning. It may have increased the influence and mobility of the origin class, which had initially been a critical determinant of SWB.

This study has some limitations. First, we must examine the impact of intra-generational class mobility on SWB. Zang and Dirk de Graaf (2016) found differences between the inter- and intra-generational class mobility effects on SWB in China. We must examine how one’s career changes affect the well-being of mobile members for a broad understanding of the relationship between class mobility and SWB in Japan.

Second, research has shown that the impact on SWB varies based on the class to and from which individuals have moved (Miwa & Yamamoto, 2012; Zhao et al., 2017). It is essential to examine how the trajectory of mobility affects the SWB of mobile members to understand the consequences of class mobility further.

Third, we must examine the origin classes of both parents. Beller (2009) insisted on the importance of the mother’s class while estimating class mobility effects. Shirahase and Ishida (2018) found that the mother’s class affects the course of individual lives in Japan. Hirao and Taromaru (2011) recommended considering unemployed part-time workers as the origin class. We did not focus on unemployed individuals. The SWB of those from an unemployed origin class is likely to differ from that of others in a rather remarkable manner. We must expand and consider the scope of the class of origin.

8 Conclusion

To date, empirical research on the consequences of class mobility and related discussions has been limited. Little research has been conducted on how class mobility produces change according to relevant macro-social contexts. Using a Japanese representative national survey, namely the SSM Survey Datasets from 1985 to 2015, we examined the impact of intergenerational social class mobility on SWB and evaluated how the relations differ based on socioeconomic factors.

We drew up the acculturation, dissociative, relative status, and inertia hypotheses based on both a literature review and empirical research. We hypothesized that the effect of social class mobility differs across cohorts whose socioeconomic situations differed. We adopted the DRM method to deconstruct the origin, destination, and mobility effects. Using this approach, we arrived at three significant results.

First, it was clear that the overall effects of intergenerational social class mobility on SWB were affected mainly by their destination class, although the origin class did contribute toward the formation of SWB. The SWB of mobile members was highly similar to that of their non-mobile destination class counterparts. In contrast, no additional mobility effects were found.

Second, the cohort analysis showed that the effects of the origin and destination classes, and mobility differed across cohorts. The SWB of the high economic growth cohort was determined only by their destination class. In contrast, although the impact of the destination class was dominant, the origin class effect became significant in the stable economic growth cohort. We confirmed that the formation of SWB in the recession cohort was remarkably different from that of the former two cohorts. The origin class impact increased considerably. Horizontal mobility had a significant negative impact. None of the hypotheses we formulated fully explained the results of horizontal mobility. One possible interpretation is that the mobility between non-manual and manual occupations that accompanies tremendous changes in occupational conditions/culture without prominent socioeconomic advantages or disadvantages may harm the SWB of mobile members. Our results show that the youngest cohort began to be significantly affected by their origin class, implying that the youngest generation became more committed to their origin class even after gaining mobility.

The gender analysis presented a similar trend in the origin and destination class and mobility effects. The effects of destination class were dominant in the first and the second cohort. The youngest cohort began to be affected considerably by their origin class. The effects of the destination class were more dominant among males. Males who began their careers during the periods of high and stable economic growth were affected by their destination rather than origin classes. We did not find the acculturation process in the two early cohorts of males.

All female cohorts were less affected by their destination class than were the male ones. Females tended to be more affected by their origin classes. We also found an additional negative horizontal mobility effect in the youngest female cohort. This shows that the effect of horizontal mobility in the youngest cohort as confirmed in the cohort analysis was attributed to females.

Overall, mobile members in Japan are currently in the course of acculturation. They have affected both origin and destination classes. However, the destination class mainly contributed toward formulating the SWB of mobile individuals. The impact of intergenerational mobility varies across cohorts, each with different socioeconomic experiences. Those who were mobile during the period of high economic growth were mainly affected by their destination classes. Those who were mobile in the recession were more under the influence of their origin classes. The results strongly show that the effects of class mobility on SWB vary across macro-social contexts. Our results imply the macro-social context should be considered while examining the impact of class mobility on SWB.