استخدام مادام العمر در ژاپن : مفاهیم و اندازه گیری

This paper addresses three questions: (1) How big is lifetime employment in Japan? (2) How unique is it? and (3) How is it changing? Through the use of multiple data sets and methods, I find that no more than 20% of workers in Japan are likely to be employed under informal lifetime employment contracts, a far smaller percentage than has been reported. Job mobility remains considerably lower in Japan than in other advanced economies (particularly the US). Evidence regarding changes in lifetime employment is mixed. The share of workers in the core is declining, but the probability of job separations has remained stable for those who are already in the system. There is also evidence that the economic stagnation of the 1990s disproportionately affected females and younger workers.

Lifetime employment has long been seen as one of the main features of the Japanese employment system. Defenders of lifetime employment have cited the many benefits that arise from stable employment relationships. Critics have pointed to the big efficiency losses associated with a labor market that cannot adapt quickly to changing demands for its products. Now, some observers say they see the system eroding – a claim that worries defenders of stable employment policy, and encourages those who push for more flexible labor markets. Yet, to date there has been little attempt to document the phenomenon systematically to reveal the extent of lifetime employment in Japan, its uniqueness in the world economy, and its rumored decline in the face of globalization. This paper addresses three fundamental questions about lifetime employment in Japan: How big is it? How unique is it? And, how is it changing? The first concerns the lifetime employment rate, which measures the proportion of the labor force that is covered by lifetime employment. Despite a number of significant contributions in the study of lifetime employment, hard data on the true extent of the practice are rarely reported. The question – “how big” is lifetime employment in Japan? – is perhaps one of the most frequently posed questions in empirical research on the Japanese labor market, yet few researchers have undertaken serious efforts to estimate the size with precision. Informed scholars know that its actual coverage is limited to a minority share of the Japanese labor force. Some point to 30% or one-third (Form, 1979, Hashimoto and Raisian, 1985 and Schregle, 1993); but, one-third of what? What is the denominator, and what is the numerator? Lack of consensus about the empirical definition of lifetime employment has caused a lot of confusion, with discussions often muddled by conflicting or overlapping concepts. The implicit nature of lifetime employment is the root of the measurement problem, as there is no unified survey method for estimating its size. Lifetime employment is not a contractual state (Itoh, 1991). The employment contract includes no explicit clause mentioning this policy, and employers are under no obligation to guarantee employment. Lifetime employment is better understood as a long-term commitment between workers and employers rather than as a permanent employment contract. As such, a worker survey cannot ask the question, “Are you covered by lifetime employment?” Likewise, an employer survey cannot ask the question, “What proportion of your workforce is covered by lifetime employment?” And even if these questions could be asked, there would be considerable discretion in their interpretations, which would introduce severe measurement bias.1 Invariably, an accurate assessment of lifetime employment requires the time dimension. Researchers need information concerning past job histories to examine patterns of job mobility, but these microdata are not readily available.2 Measuring lifetime employment in Japan therefore requires simplifications and approximations from which we may deduce its size and speculate on its direction of change. I review the existing literature and methods used to estimate the size of lifetime employment, evaluate their strengths and weaknesses, and present the latest available measures. I conclude that lifetime employment applies to about 20% of the working population in Japan, a much smaller share compared to those previously reported. While estimations of lifetime employment improve our understanding of the Japanese labor market, they are more valuable if we know how this picture compares to other economies. The second question assesses the uniqueness of Japan’s long-term employment in an international context. One of the problems underlying comparative analysis of labor markets is insufficient data. Consequently, much of the existing research relies on US–Japan comparisons. However, these two countries represent two extremes in job mobility, so the findings show unanimous support for the resilience of long-term employment in Japan when compared to the US. Despite the data constraints, I attempt a comparative analysis to evaluate the uniqueness of lifetime employment in Japan relative to the US and other countries. My third question looks into changes in the lifetime employment practice over time. This area has received the most attention, perhaps more so in policy and the media than in the academic community. The debate over the demise of lifetime employment is hardly new. It has been a recurring and evolving theme during the postwar period, partly in response to fluctuations in the business cycle, not only during the slump years but also during the growth years (Moriguchi and Ono, 2006 and Ono, 1997). Most recently, the debate focuses on the “lost decade” of the 1990s and its impact on the lifetime employment system. If we take the journalistic coverage at face value, then all signs point to the end of lifetime employment in Japan. But on what basis can we make this claim? How we observe changes in lifetime employment really depends on how we define and measure it, which puts more emphasis on our first question. Indeed, for every scholarly article written about the demise of lifetime employment, there is another which vouching for its stability and resilience.3 In parallel to this debate in Japan, a panel of experts examined trends in job stability and job security in the US, primarily in response to the journalistic and anecdotal accounts suggesting the end of “lifetime jobs” in the US in the 1990s. Their work (published in an edited volume by Neumark (2000)) found some evidence of declining job stability in the 1990s, but concluded that these changes were not very large, and have not persisted long enough to constitute a widespread shift in the US labor market. The absence of large, high-quality datasets on Japan like those available in the US, make it difficult to carry out a comprehensive and systematic evaluation of the state of employment stability in Japan. Still, we can track and monitor movements if we define the concepts and measurements of lifetime employment clearly and consistently. In some respects my analysis leads to seemingly conflicting findings. The final section of the paper attempts to make sense of these in a consistent framework. Thanks to the keen interest in the academic community, there is now a rich and extensive theoretical and empirical literature on all possible aspects of lifetime employment in Japan. I contribute to the existing literature by providing a systematic and comprehensive assessment of the size and scope of lifetime employment, an undertaking which has been largely overlooked in previous research. By documenting its size with precision and clarity, I hope to lay out the framework for more accurately assessing the uniqueness and changes in the lifetime employment system. As a further contribution, I look into gender differences in lifetime employment. The lifetime employment system is gender biased by design (Ono, 2007). It favors men, who are moreable to make long-term commitments, and disfavors women who are less likely to do so because they are expected take on family obligations. However, it is hasty to assume that women are automatically excluded from coverage, which is the underlying assumption in many of the earlier studies. I examine the data for both men and women, and evaluate the extent of gender asymmetries in the lifetime employment system.

How big is lifetime employment in Japan? While lifetime employment is often touted as one of the main pillars of the Japanese employment system, in practice it applies only to a small portion of the labor force. My estimations based on various methods point to roughly 20%, a much smaller percentage compared to previous estimates: (i) The ex-ante measure of the core workforce, which assumes that lifetime employment coverage is limited to male standard workers in large firms (⩾500) and in government, is 19%; (ii) The proportion of lifetime workers defined as those in the age group 50–54 who have never left their employers since school graduation, is 20%; and (iii) The probability of surviving job separations for 30 years is 20%. Yet, the lifetime employment rate varies considerably by gender, firm size, and education level. My analysis solidly confirms the generalized view that female workers do not benefit from the lifetime employment system, while male workers in large firms (and in government) are most likely to be covered by it. At its extreme, the proportion of lifetime workers among male university graduates in large firms is 46%, and the 30 year survival probability of male workers in government is 65%. How unique is lifetime employment in Japan? If lifetime employment is taken as a generalized practice of long-term employment, then the practice is more pervasive in Japan than in other countries. Job mobility in Japan remains one of the lowest among the OECD member countries, as characterized by the following features: (i) the highest proportion of workers with tenure longer than 20 years; (ii) lowest proportion of workers with tenure less than one year; (iii) highest 5 year retention rate; (iv) lowest separation rate; and (v) lowest involuntary separation rate. At the opposite end of the scale lies the US, characterized by high job mobility. By any measure, job mobility is much higher in the US than in Japan. Comparison of job mobility between these two countries should thus be understood as a comparison of two extremes. And, how has lifetime employment changed over time? Understanding trends in lifetime employment requires a multi-dimensional approach. We must examine a wide range of data measures, and not one in isolation. I present a summary table of the various estimations in Table 7. Changes in lifetime employment are shown under the column “lifetime employment” where the (+) sign indicates support for it and the (−) sign indicates its decline. The results are mixed. Overall, the ex-ante measures – the core workforce, standard employment, and full-time employment (as the share of total employment) – suggest a decline in lifetime employment. On the other hand, the ex-post measures – the proportion of lifetime workers, and the probability of surviving job separations – show an upward trend or no change. The two are not necessarily conflicting trends but require some elaboration. Much of the disagreement stems from confounding or not disaggregating the inflow, outflow and the stock of workers.Although the population of workers who are ex-ante covered by lifetime employment may be shrinking, the likelihood of job separations has remained stable for those who are already in the system. Consider the analogy between the labor force and the bathtub where the water flowing into the tub represents the flow of workers into the core, water flowing out is the flow of workers out of the core, and the water level in the tub is the stock of the core workforce. In this analogy, the water level remains the same or decreases because: (i) There is little water being eliminated from the tub; and, (ii) Preserving the current water level requires choking off the flow of water into the tub. The mobility measures add support to this effect. In the 1990s, new graduates were significantly less likely to enter the labor force as standard workers than in previous periods. This led to two outcomes. First, standard employment declined relative to nonstandard employment, which led to an overall reduction of the core workforce. And second, the expansion of the nonstandard workforce among younger workers resulted in their higher job mobility, as characterized by their lower retention rates and higher separation rates. In contrast, job mobility among older workers remained virtually unchanged during the 1990s (with the exception of the post-retirement age group). In a sense, we can surmise that this is a uniquely Japanese response to globalization. The Japanese employment system has always been criticized for its rigidity and inability to respond quickly to business cycle fluctuations. In order to be more flexible and responsive to changes in the global economy, companies are reducing their core, and expanding their periphery labor force. But companies, especially the large ones, are still honoring the implicit contract of lifetime employment and protect those who are in the core. Hence simplified statements that decry the end of lifetime employment are overblown, if not misplaced. Lifetime employment is far from dead for those who are in the core. The ongoing debate concerns whether this outcome is attributable to changes in the supply side or the demand side. The consensus is shifting towards the latter. An increase in involuntary job separations accompanied by a decrease in voluntary separations suggests that job mobility was affected by the economic conditions of the 1990s. Outside employment prospects for those who were already in the core were not favorable, so the core workers held onto their jobs. The firms responded by avoiding dismissals to the fullest extent possible, and instead cut back on recruitment of new graduates. Genda (2001) claims that the adverse conditions of youth employment in the 1990s were the consequence of an overemphasis on the employment protection of middle- to older workers. The economic downturn triggered a hiring freeze where employers prioritized their core workers at the expense of new hires. The data show that many of the new graduates became freeters involuntarily; they would have preferred standard employment, but were unable to do get it because of unfavorable demand-side conditions.22 Meanwhile, survey results from both workers and employers indicate strong support for the lifetime employment system. In fact, the support base seems to be increasing. A worker survey conducted in 2004 shows that 78% of workers are in favor of lifetime employment, an increase of 5.7% from the first survey conducted in 1999 (JILPT, 2005b).23 Older workers are more likely to support the system, presumably because they are more likely to benefit from it, but the majority of the younger workers also show support for it (64.2% of males and 66.4% of females in their 20s). On the employer’s side, survey results from 1999 show that 60.6% of personnel managers intend to “maintain lifetime employment as much as possible,” an increase from the 46.8% reported in 1997.24 These survey results are strong counterevidence to the anecdotal view that lifetime employment is a thing of the past. Workers – both the young and the old – desire employment security, and employers still feel obliged to preserve the system. But lifetime employment is never without its critics, and the criticism mainly concerns its lack of flexibility in adapting to ongoing changes in the business environment. The most recent of these relates to the difficulties of achieving work and family balance. Since its inception, the lifetime employment system presumed a complete specialization between the sexes with one spouse fully committed to market work and the other devoted to non-market work. Employers in turn supported this practice by paying workers high enough wages so that spouses did not have to work (Moriguchi and Ono, 2006). As such, these underlying assumptions are outdated and incompatible with the trend toward equal participation between the sexes. Women, who are on average not in the position to make the same form of long-term commitments with employers as men do, have been systematically excluded from the lifetime employment system. Women who desire professional careers are forced to decide whether to work or to start a family. Against the backdrop of the alarming decline in fertility rates in Japan, the government ministries have been trying to remove some of the constraints of the rigid employment system, e.g. by requiring employers to install parental leave programs. However, these efforts have not made much headway so far. Among men, less than 1% took parental leave in 2005 (see also the low percentage of men who left their jobs for reasons related to “marriage, childbearing and homecare” in Table 6). The top reasons they cited for not taking leave were “because the work environment is not conducive to taking leave,” and “because it may affect my promotion” (MHLW statistics). The falling birth rate and the rigid employment system in fact are intricately linked. It may well require a paradigm shift in the cultured norms and mentality among Japanese employers and workers in order to achieve a better balance between work and family in the long run. One of the big advantages of the exercises that I have presented here (with the possible exception of survival analysis) is that they can be easily replicated and updated. The data are taken from aggregate statistics published by the Japanese government and other publicly available sources. In documenting the basic concepts and measurements of lifetime employment, I hope to have set out the basic tools and the framework for more accurately gauging the size, uniqueness and changes in the lifetime employment system.