2004/11 Research & Review

The State and Change in the "Lifetime Employment" in Japan: From the End of War Through 1995

YAMAGUCHI Kazuo
Visiting Fellow, RIETI

Introduction

There are two ways of interpreting "lifetime employment": the rational and functional interpretation, and the historical and cultural interpretation. According to the former, lifetime employment can be understood to be a system to reduce outflow of employees with firm-specific human capital - those with knowledge and technique acquired through training and experience in the firm - by giving them special opportunities for promotion and offering premium remunerations when there is a large value placed on such firm-specific human capital. Similar to the retirement allowance system, the seniority wage system can be understood as a method deferring payment of wages, which provides employees with incentives to remain employed by a firm for an extended period of time by setting wages below marginal productivity for those who have been employed for a relatively short period, while offering wages greater than marginal productivity for those who have been employed for a longer period. However, this system inevitably raised personnel costs in firms in regards to maintaining senior employees, and necessitated the introduction of a mandatory retirement system at a relatively young age.

An account of the lifetime employment system based on historical and cultural interpretation can be seen in "Bunmei to shiteno Ie Shakai (The 'Ie' Society as a Civilization)," Murakami, Kumon and Sato, 1979). The authors view lifetime employment system as a "kintract," the combination of kinship and contract, that has brought certain elements of the feudal "Ie (household)" system into modern society. They consider loyalty to the "Ie" household as also being applicable to modern-day firms. However, their theory falls short in terms of providing explanations for social change regarding when and why a particular employment system develops and declines, and thus is inferior to the rational and functional interpretation.

The lifetime employment system is subject to various theoretical interpretations. However, in order to understand this system analytically, it is necessary to distinguish the following two elements:

Element 1: The lifetime employment system as the employer's policy - a system in which the employer refrains from dismissal or lay-offs of full-time regular employees for management reasons.

Element 2: The lifetime employment system as the employees' commitment - to not leave or switch employment until they reach retirement age once they have secured a full-time regular position with their employer.

Studies on the lifetime employment system differ in content depending on which of the above two elements researchers focus on, although the two kinds of studies are complementary. In studies focusing on the element 1, the firm is the unit of survey research and the lifetime employment system is regarded as an aspect of broadly defined employment adjustment by firms. Naturally, characteristics of firms are used as explanatory variables in analysis. On the other hand, the individual employee is the unit of survey research in studies focusing on element 2. With an exception of some employer information obtained from employees, employees' individual attributes such as type of employment are used as explanatory variables in the analysis. From the viewpoint of element 1, it is possible to define a modified concept of lifetime employment by hypothesizing that employees' "permanent" commitment to their employers begins at a certain age (around the age of 30) after experiencing "re-matching" with several jobs in the labor market. Naturally, it is accompanied by a study of re-matching between personnel and job in the labor market, which is a macro-level employment adjustment.

I analyzed the lifetime employment system from the viewpoint of element 2 and how it changed during the time from the early post-war period to 1995, focusing on male employees who obtained their first full-time job after the end of World War II. The Japanese economy has deteriorated since 1995, and it is assumed that the state of employment has gone through drastic changes in the process. Nevertheless, it is important to reexamine what changes occurred in the lifetime employment system over half a century prior to this period, during which time Japan had gone through several economic stages including: postwar chaos, the period of high economic growth, the subsequent period of stable growth, and the swell and burst of the bubble.

Theoretical viewpoints and hypotheses

Basic requirements to keep lifetime employment viable are that 1) employers expect that their demand in the workforce - especially for firm-specific human capital - will grow or remain unchanged in the future, and that 2) benefits gained from securing employees' long-term commitment (or the costs of labor turnover) are greater than benefits gained from maintaining flexibility in terms of adjustment of the employee population. We thus obtain the following hypotheses.

Hypothesis 1: Relatively large firms that find significant benefits in nurturing firm-specific human capital are more inclined to adopt a lifetime employment policy for full-time regular employees. Accordingly, the probability of employees' lifetime commitment to their employers will increase with firm size.

Hypothesis 2: Employers tend to apply a seniority-based wage system to employees with jobs that are likely to yield firm-specific human capital, and employees with such jobs tend to have a high probability of lifetime commitment to their employers than employees with other jobs unless they quit or change jobs at an early tenure.

Hypothesis 3: Those who obtained their first full-time employment at a time when demand for workforce was expanding will be less likely than others to quit the job, and thus their years of employment and the probability of lifetime commitment to their employers will be greater.

Generally, labor mobility is greater for blue-collar workers than for white-collar workers. For large Japanese firms however, the wage system of blue collar workers resembles that of white collar workers (Kazuo Koike, Shigoto no Keizaigaku (The Economics of Work, 1991). We can expect a similar resemblance of blue-collar workers to white collar workers concerning the probability of lifetime commitment to the employer among employees of large firms.

Hypothesis 4: Although blue-collar and white-collar employees of large firms are expected to show the same level of (relatively high) lifetime employment probability, this tendency will not be found for employees of small and medium-sized firms.

The seniority wage system, a method of increasing the incentive for employees to make a long-term commitment to their firm, was adopted more prominently in the public sector than in the private sector. Under such a system, the amount of wage that employees must forfeit by leaving employment increases with years of seniority. Thus, if an employee under the seniority wage system is to quit his job, it will be rational for him to do so at an early stage of tenure.

Hypothesis 5: Provided that they are to leave their employer, government employees will do so at an earlier tenure compared with employees of large private-sector firms.

Being unresponsive to change in the market situation, government agencies will be more prone to institutional inertia. That is, institutions established within the public sector are less likely to change than those established within the private sector. Thus, if economic conditions change in the direction to weaken the foundation for lifetime employment, the gap between the public and private sectors will likely widen in terms of the degree to which the lifetime employment system is retained.

Hypothesis 6: In recent years that witnessed a substantial rise in youth labor mobility, a gap between the public and private sectors regarding the degree of employees' lifetime commitment to their employer has widened.

Individual talents and jobs may not be matched well in initial employment, thereby necessitating an ensuing period of re-matching between personnel and job in the labor market. Normally, such matching adjustments are caused by employers firing disqualified employees and by employees leaving employment voluntarily. However, in cases where the lifetime employment policy of employers is prevalent, the adjustments of the latter pattern will likely be dominant. Employees are expected to make a lifetime commitment to their employer with a greater probability after such a period of matching adjustments in comparison to at the time of their initial employment. The disparity in the probabilities of employees' lifetime employment among different occupations and among firms of different sizes will diminish after the period of matching adjustments.

Hypothesis 7: The probability of not changing or leaving the job from age 30 to mandatory retirement will be substantially higher than the corresponding probability for the initial employment.

Hypothesis 8: Differences in probability among employees of different firm sizes and different occupations of not changing or leaving their jobs until mandatory retirement will decrease for jobs held at age 30 compared with initial jobs

In the U.S. and Europe, matching adjustments in the labor market for young people are considered to be a career development process in which those with relatively extensive educational background try to obtain employment that meets their credentials and qualifications, as well as a professionalization process in which individuals try to obtain re-education or appropriate professional qualifications after recognizing the disadvantage of lacking these elements. Although the function of professionalization during the period of re-matching between personnel and job has not been well-recognized in Japan, the following hypothesis can be tested.

Hypothesis 9: Among people who change their jobs before age 30, the proportion of people with professional occupation will be greater at age 30 than at the time of their initial employment.

To verify the above hypotheses, we employ job history data based on national representative sample of men taken from the 1975, 1985, and 1995 Social Stratification and Social Mobility Survey in Japan. In addition to the statistical regression models for the hazard rate of job leave, we also use a specialized regression model for hazard rate (Yamaguchi, Journal of the American Statistical Association 87, 1991) that distinguishes between determinants of employment duration and determinants of probability of lifetime employment (the probability of not leaving the job until mandatory retirement). The main results of the analyses are as follows:

Analysis results

1. On the rate of job leave and the probability of lifetime employment from the time of initial full-time employment:
(1) Differences by firm size and occupation in hazard rates of leaving the first full-time employment occur primarily due to differences in lifetime employment probability, though also partly due to differences in years of employment duration. This trend remains unchanged independent of time period.

(2) Among men employed in the postwar period, the probability of lifetime employment was 40 percent or above for employees of large firms and government agencies regardless of type of occupation, and for professional/technical and clerical employees at small and medium-sized firms. This trend was most conspicuous for the third generation of postwar employees (born in the 1940s).

(3) The resemblance of blue-collar workers to white-collar workers among employees of large firms can be observed in the pattern of employment. High probability of lifetime employment for blue-collar workers is a characteristic unique to employees of large firms, and this tendency remains unchanged independent of time period.

(4) An increase in the hazard rate of leaving their jobs for relatively young fourth-generation postwar employees (born in the 1950s onward) is more attributable to the shorter period of time they remain at their first job rather than to a decrease in the expected probability of lifetime employment.

(5) Among the fourth-generation postwar employees who are relatively younger than people of other generations under study, the probability of lifetime employment has substantially increased for government employees whereas the probability has decreased somewhat for employees of large private-sector firms. This resulted in the remarkable widening in recent years of the gap between government employees and employees of large private-sector firms in terms of probability of lifetime employment.

(6) Government employees who leave their employer tend to do so at a relatively early stage, apparently because the seniority-based wage profile has a steeper slope of increase with age for government employees than for employees of large private-sector firms.

2. On changes in jobs between the time of the initial full-time employment and the employment held at age 30:
(1) Among those who were initially employed full-time after the war, about 40 percent had retained their first job at age 30, 50 percent were employed in a different full-time job, and the remaining 10 percent were not employed full-time (many of them had become self-employed).

(2) The rate of leaving a full-time employment before age 30 is particularly high for those working for small firms with 30 employees or less, those with a sales occupation, and the first-generation postwar employees (born in the 1920s).

(3) Among those who were employed full-time at age 30, the odds of leaving the initial job before the age of 30 is particularly high for employees of small and medium-sized firms, and those without a professional/technical occupation.

(4) Among those who were employed full-time at age 30, the odds of leaving the initial job before age 30 are higher for younger cohort members. This indicates that the re-matching between personnel and job prompted by employees' voluntary job changes has become more prevalent in recent years.

(5) Among those who have changed jobs before the age of 30, 54 percent moved between firms of different sizes (based on the six-category classification of: firms with 5 or fewer employees, those with 6 to 299 employees, those with 300 to 999 employees, private-sector firms with 1,000 or more employees, and government agencies). However, no significant differences are observed in the distribution of jobs by firm size between the initial jobs and jobs held at age 30. This indicates that there is no barrier in mobility into employment at a larger firm or a government agency before the age of 30.

(6) On the other hand, among those who changed jobs before the age of 30, 36% moved between different occupations (based on the four-category classification of: professional/technical occupations, administrative/clerical occupations, sales occupations, and manual occupations). The occupational distribution differs significantly between the initial job and the job held at age 30. In particular, the proportion of those holding a professional/technical job at age 30 is 1.7 times as large as this proportion at the initial job, demonstrating a conspicuous trend toward professionalization.

(7) An analysis of determinants of the odds of professionalization among those whose first job is not a professional/technical occupation shows that there exists no greater barrier to obtaining a professional/technical job by age 30 among those initially employed in small and medium-sized firms compared with those initially employed by government agencies or large firms. However, graduates from either a four-year college or from a junior technical college have higher odds of obtaining a professional/technical occupation than others. Controlling for education, those who initially held a manual job have smaller odds of obtaining a professional/technical occupation than others by age 30.

3. On the rate of job leave and the probability of lifetime employment from age 30
(1) As is the case for the initial full-time employment, the hazard rate of leaving the full-time job held at the age of 30 is greater for those working for small or medium-sized firms than for those working for a large firm or government agency, and is also greater for those who hold a sales or manual occupation than for those who hold a professional/technical or administrative/clerical occupation.

(2) The first distinctive feature of the hazard rate of job leave for the initial full-time job held at aged 30 as compared to the rate for the initial full-time job is a significantly lower rate for employees of large private-sector firms compared to that for employees of government agencies. Thus, the gap in the hazard rate of job leave between the public sector and the private sector is greater after age 30.

(3) The second distinctive feature of the hazard rate of employment leave for the full-time job held at age 30 as compared to the rate for the initial full-time job is that the cohort differences observed in the hazard rates of leaving the initial full-time job (a higher hazard rate for fourth-generation postwar employees born in the 1950s than for older cohorts) are not significant for the hazard rates of leaving the full-time job held at age 30. This indicates that, while the tendency for leaving full-time job during the period before the age of 30 becomes greater for younger cohorts, such a change does not exist for the rate of job leave after the age of 30.

(4) The third distinctive feature is that the interaction effect of firm size and the distinction between blue-collar and white-collar workers is not present for the hazard rate of leaving the full-time job held at age 30. This indicates that the relatively low rate of job leave among blue-collar workers in large firms, comparable to the rate among white collar workers in those firms, is attributable to characteristics of the rate of job leave before the age of 30 rather than to the characteristic at age 30 or later.

(5) Those who work at age 30 for a firm larger in size than their initial employer, have a greater hazard rate of leaving their jobs than job changers. This seems to suggest that the work environment provided by relatively large firms is less likely to offer job security for job changers from smaller firms than for other job changers.

(6) The probability of retaining the same employer for 25 years from age 30 was maintained at a very high level of about 50 percent throughout the postwar period regardless of cohorts. This shows that full-time workers' lifetime commitment to a single employer - except for the period of matching adjustment in the labor market before the age of 30 - has firmly taken a root as a broadly accepted practice in postwar Japan.

(7) The probability of 25-year retention of the same employer for the job held at age 30 was in the range of 50-55% among white-collar workers of private-sector firms with little difference by firm size, whereas the probability for government employees was larger and above 70 percent.

(8) On average, the probability of 25-year retention of the same employer for the job held at age 30 among blue-collar workers is lower than that of white-collar workers. However, when the two groups of workers are compared in terms of change in the probability of lifetime employment from that expected at their initial job to that expected at age 30, an increase is more conspicuous for blue-collar workers than for white-collar workers. This means that the adjustment in the labor market through re-matching of personnel and job before the age of 30 plays an important role in enhancing job security for blue-collar workers.

On Issues of the Current State and the Near future

Japan's employment practices have allegedly changed drastically since 1995, rocked by the wave of corporate "restructuring." However, specific episodes aside, no great change has been taking place in terms of average duration of employment. Thus, we should not prematurely conclude about "drastic changes." However, I believe that the following points are important:

(1) Japanese firms have been increasingly dependent on contract employees and workers dispatched by personnel agencies for labor while reducing the number of full-time regular employees. As a result, it is expected that the probability of lifetime employment of full-time employees aged 30 or older will differ greatly depending on whether their employment status is regular or non-regular, thus increasing the likelihood for bipolarization of full-time employees between those with job security and those without.

(2) In relation to the point discussed in (1), it is increasingly becoming important to view employment security not only by the duration of employment for the same employer, but also by the continuation of employment in the same type of job or in occupational career regardless of employer change.

(3) The matching adjustment between personnel and job before the age of 30 has become more important. However, there have recently been increases in the rate of youth unemployment, in the proportion of young people with temporary jobs (so-called "freeters" in Japan), and in the number of those who are "not in education, employment or training (NEET)." As a result, the average period of time that post-school youths are unemployed or with temporary jobs has become substantially longer, thereby possibly causing inefficiency in the pre-30 matching adjustment in the labor market which also affects job security for jobs held at age 30 onward. It is important to clarify this possible trend while at the same time promoting professionalization, which is an important function for facilitating re-matching between personnel and job. The growing NEET phenomenon (see "NEET: Freeter demo naku Shitugyosha demo naku (NEET: neither being a freeter nor unemployed)," Genda and Maganuma, 2004), also seen in Britain, is the problem of entry-level barriers that shut the door to young people even before they experience re-matching. Identifying the causes of this phenomenon is a task of utmost urgency.

(4) Based on the existing wage system, the planned postponement of mandatory retirement age to 65 would result in a substantial increase in personnel costs. In order to avoid this, Japanese firms are expected to make significant changes in their wage systems. There is a significant possibility that such moves will lead to a major change in employees' commitment to their employers. Given such a prospect, it is important to closely monitor the rate of job leave at age 30 or later, and examine how it is related to changes in the corporate wage system.

>> Original text in Japanese

March 11, 2004