Japan: Selected Issues
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

This Selected Issues paper analyzes Japan’s medium- and long-term fiscal challenges. It discusses the initiatives that will be necessary to cope with Japan’s medium- and long-term fiscal challenges. It describes medium- and long-term projections of the fiscal position, and assesses the size of the consolidation measures needed to restore long-term fiscal sustainability. The paper explores possible options for consolidation and offers a representative package of such measures. The paper also describes the projected baseline path for the long-term fiscal balance, incorporating the long-term pension reform plan formulated in 1994.

Abstract

This Selected Issues paper analyzes Japan’s medium- and long-term fiscal challenges. It discusses the initiatives that will be necessary to cope with Japan’s medium- and long-term fiscal challenges. It describes medium- and long-term projections of the fiscal position, and assesses the size of the consolidation measures needed to restore long-term fiscal sustainability. The paper explores possible options for consolidation and offers a representative package of such measures. The paper also describes the projected baseline path for the long-term fiscal balance, incorporating the long-term pension reform plan formulated in 1994.

II. Japanese Labor Market Developments 1/

1. Introduction

Japan is emerging from its worst recession in the post-war period. At the same time, its population is evolving from one of the youngest to the oldest among developed economies. These developments have placed severe strains on its labor market, with the unemployment rate setting postwar highs. While the resumption of growth will undoubtedly ease labor market pressures, continued demographic shifts, as well as structural changes in the Japanese economy, should present further challenges over the medium term.

While aware of the dangers of oversimplification, the Japanese labor market can be characterized as a system in which there has been relatively high job security (as reflected in low unemployment rates by international standards, with lifetime employment attachments for a significant proportion of workers) and steep wage-tenure profiles. Labor flexibility has been achieved through significant variation in average working hours, including overtime; flexible and synchronized wage determination (including bonus payments); and, to a lesser extent, flexible employment and participation rates, especially among female workers.

This chapter examines developments in Japan’s labor market in the face of recent economic and demographic pressures and some policy issues that could be addressed to facilitate the efficient operation and evolution of the labor market

2. The “lifetime” employment system

In examining recent developments in Japan’s labor market, it is important at the outset to understand the role played by Japan’s much lauded “lifetime” employment system. 2/ Under this system, workers are generally hired directly after graduation. Firms implicitly agree not to lay off or fire workers, while workers accede to flexible work assignments, including, if need be, physical relocation to another plant or office of the firm, or even to an affiliated firm. The firm and the employee invest heavily in training, a good deal of which may be firm specific. Wage-tenure profiles are steeply tilted by international standards, providing incentives for the firm to invest in employees’ human capital, and incentives for employees to remain with the firm.

Hashimoto and Raisian (1985) provide evidence that job retention rates are considerably higher in Japan than in the United States. In addition, while job “shopping” is greater in both countries in early working years, it is relatively less frequent in Japan, and job turnover during the lifecycle is much greater in the United States. 1/ Data indicate that this aspect is not limited to large firms alone, as average male job tenures in Japan exceed their United States counterparts, by an even greater degree for smaller than larger firms. 2/ Thus, the assertion that small firms act as a “buffer” for labor market pressures in Japan does not appear to be entirely accurate.

Additional stylized facts concerning the lifetime employment system are steep wage-tenure profiles and low turnover rates. As the data in Table II.1 indicate, wage tenure profiles are far steeper in Japan than in Western Europe. 3/ Hashimoto and Raisian (1989) compare United States and Japanese practices, finding that, after controlling for education, wage-tenure profiles are steeper in Japan than in the United States across all firm sizes. 4/ In both countries, however, wage-tenure profiles are steeper at larger firms, suggesting that they undertake greater worker investments. Mincer and Higuchi (1988) examine wage-tenure profiles and turnover rates in Japan and the United States using micro-data, and conclude that as much as two-thirds of differentials in turnover rates are explained by differences in the steepness of wage profiles.

Table II.1.

Japan: Wage Levels and Worker Distribution by Length of Service

(Wages for workers with less than 2 years of continuous service = 100)

article image
Sources: Japan: Basic Survey on Wages Structure; A as of 1976; B as of 1990; C as of 1994. Other countries: 1972 Structure of Earnings in Industry; reproduced from Tokyo Financial Review (Bank of Tokyo) Vol. 20, No.11, November 1995. Notes: in A and B * denotes 2 years or less, ** 3-4 years, and *** 30-44 years. Workers wages are monthly ordinary wages in Japan, hourly wage for production worker in EC and average monthly salary for clerical, managerial and technical workers in EC.

How widespread is the lifetime employment system? Given the evidence that long job-tenure and relatively steep wage-tenure profiles characterize Japanese employment patterns for male workers in firms of all sizes, but are more pronounced among larger firms, a lower bound on the extent of lifetime employment may be approximated by the proportion of male employees in large firms. In 1994, 21 percent of male employees worked in firms employing 1,000 workers or more. 5/ However, given that smaller firms also have longer tenures in Japan than in the United States, it may be appropriate to consider a wider scope for this measure. In this case, over 43 percent of male employees worked in firms with more than 100 employees. Thus, to the degree that smaller firms emulate the employment practices of larger firms, lifetime employment is an important feature of the Japanese labor force. 1/ 2/

3. Population, labor force and participation rates

The Japanese economy experienced what has been generally characterized as a “bubble” period of unsustainable economic growth and asset price inflation in the late 1980s, followed by its worst postwar cyclical downturn. 3/ Real output grew by an annual average of 4.8 percent during 1987-91, followed by average annual growth of only 0.7 percent during 1992-95. Japanese labor markets underwent a similar cyclical pattern, with labor force and employment growth significantly outpacing growth in the working age population in the late 1980s, followed by a subsequent drop-off (Chart II.1). In fact, the working-age population actually declined in 1995, and under current demographic projections, is anticipated to shrink further for the foreseeable future. 4/ Therefore, any future increases in the labor force would be the result of increased participation rates of the working age population, or postponed retirement ages. As seen in Chart II.2, the participation rate did increase during the bubble period, in part in anticipation of future demographically-induced labor market shortages, but has declined somewhat subsequently. 5/

CHART II.1
CHART II.1

JAPAN: WORKING AGE POPULATION, LABOR FORCE AND EMPLOYMENT GROWTH, 1960-95

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

Shaded areas indicate recessions as defined by the EPA.
CHART II.2
CHART II.2

JAPAN: LABOR FORCE PARTICIPATION RATES, 1960-95

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

Shaded areas indicate recessions as defined by the EPA.

One interesting facet of labor force participation rates is the changing patterns of male and female participation. Both rates declined steadily throughout the 1960s and early 1970s, reflecting declining shares of self- and family-employed agricultural workers. The male rate continued to decline until about the mid-1980s, with no discernable trend thereafter. The female participation rate, in contrast, has been on an increasing trend since the mid-1970s. Among the reasons given for higher female labor force participation are: an increase in the shares of part-time and service-sector employment which are more heavily represented by female workers; and increasingly positive attitudes of females toward work. 1/

4. Employment, hours worked and nonregular employment

Japanese employment levels are stable by international standards. This is in part a reflection of smaller fluctuations in Japanese output. Another factor is the continued high, albeit declining share of self-employed and family workers. In 1989, these workers accounted for almost 19 percent of nonagricultural employment, compared to about 8 percent in the United States, and still totaled 14 1/2 percent in 1995. Even after accounting for this, however, the relative stability of employment relations for lifetime employees differentiates Japan from other developed economies.

Hashimoto (1993) compares United States and Japanese labor market adjustments in the face of changing demand for output. He found that there was less adjustment in employment among Japanese manufacturing industries than in their United States counterparts, while demand shocks were buffered more by adjustment in Japanese inventory levels. 2/ He also found that Japanese adjustment patterns differed significantly before and after 1975, the year the Employment Insurance Law was adopted. This legislation supplemented traditional unemployment compensation provided directly to workers with a system in which subsidies are provided to employers, who in turn provide compensation to furloughed workers. Moreover, subsidies are provided when employers implement short-time schedules. Hashimoto found evidence that this incentive for employers to adjust working hours rather than employment resulted in reduced employment volatility, while increasing volatility in hours worked.

Abraham and Houseman (1993) also examined employment and working hours adjustment patterns in Japanese and United States manufacturing firms. They confirmed that Japanese workers enjoy greater employment stability, while the adjustment of average hours is about the same in the two countries. 1/ The authors also examined differences in adjustment patterns within each country. While they find that production workers bore a significantly greater share of adjustment than nonproduction workers in both countries, the differences between the two was much less in Japan. Indeed, they conclude that “… production workers in Japan enjoy a degree of employment stability that is similar to that enjoyed by nonproduction workers in the United States.” 2/ When comparing differences between male and female adjustment patterns, they found that while female employment and hours worked adjust significantly more than their male counterparts in both countries, the differences are greater in Japan. Finally, they compared employment adjustment patterns by firm size in Japan, and found that in most cases there was no significant differences in employment adjustment by different size firms.

How have employment and hours worked behaved over the recent business cycle? Chart II.3 displays data on employment and the unemployment rate for the present and previous three business cycles. The present cycle has seen the weakest employment performance of any of the past four recovery periods, with no measurable gains in either overall or regular employment. Regular manufacturing employment has fallen almost continually since its historical peak in mid-1992, with a cumulative loss of 570,000 workers (equivalent to 5.6 percent of initial employment). At the same time, the unemployment rate has risen to 3.4 percent of the labor force, with 2.3 million unemployed in April 1996.

CHART II.3
CHART II.3

JAPAN: COMPARISON OF FOUR CYCLICAL DOWNTURNS 1/

Index (trough of cycle=100)

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

Source: Nikkei Telecom.1/ Troughs defined as: 1993Q4 for current downturn; 1987Q2 for 1986-87 downturn; 19B3Q2 for 1982-83 downturn; 1975Q1 for 1974-75 downturn.

As seen in Chart II.4, average hours worked had also fallen to historical lows in the last recession. The secular decline in scheduled hours worked since 1988 may have been in part the result of the Labor Standards Law, which targeted a level of 1,800 hours per year by 1996. During the economic boom in the late 1980s, declining scheduled hours worked was partly offset by an increase in overtime hours. However, the ensuing recession saw a sharp drop in overtime hours, while scheduled hours worked continued to decline. 3/ Overtime hours finally increased in 1995, reflecting a hesitancy among employers to increase hiring, preferring instead to increase overtime hours. 1/

CHART II.4
CHART II.4

JAPAN: HOURS WORKED, 1970-95 1/

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

Shaded areas indicate recessions as defined by the EPA.1/ Data cover firms with 30 or more employees.

Clearly, patterns of employment and hours worked patterns have been significantly affected by the recent recession. But the recession itself has been, by most measures, the most severe in the postwar period. The relevant question is whether labor market developments have been out of line in relation to economic activity compared to previous recessions. Chart II.5 provides some basis for arguing that the sharp drop in hours worked and the decline in manufacturing employment have in fact been somewhat larger than would have been expected from past patterns. 2/ The overprediction during 1988-94 for both overall and manufacturing hours worked probably reflects the Labor Standards Law’s targeted reduction in hours worked. Overall employment changes were not out of line with past relationships with real GDP while data for 1995 suggest that manufacturers have adjusted employment somewhat more than expected.

CHART II.5
CHART II.5

JAPAN: EMPLOYMENT AND HOURS WORKED PREDICTION ERRORS, 1974-95

(Four-quarter moving average)

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

As noted above, firms are reluctant to lay off or fire their full-time regular workers during economic downturns, given investments in firm- specific human capital. In addition to making flexible use of working hours, firms have also increasingly relied on adjustments in employment levels of part-time regular and nonregular employees. For example, the share of nonagricultural employees working less than 35 hours rose from 11 percent in 1985 to 19 percent in 1994, before falling back to 17 1/2 percent in 1995. Temporary workers (those employed with contracts lasting more than one month but less than one year) and daily workers have accounted for about 10 percent of nonagricultural employees over the last decade, with very little variation.

While the majority of working women engage in regular employment, they also account for the overwhelming share of nonfull-time employment. Houseman and Osawa (1995) report that, based on the 1992 Employment Status Survey, women comprised 95 percent of all part-time workers, and 66 percent of temporary and day workers. 1/ Females comprised 49 percent of arubaito workers, which accounted for 5 percent of total employees, and accounted for 70 percent of dispatched workers (i.e., those provided by temporary employment agencies), although due to legal restrictions the latter accounted for less than 1/2 percent of all employees in 1992. The rate of part-time employment is particularly high in wholesale and retail trade (28 percent of total sectoral employment in 1992), services (17 percent), and manufacturing (14 percent). Temporary employment is also important in these sectors, and even more so in seasonal activities such as agriculture, fisheries and construction. Houseman and Osawa report data from the Ministry of Labor Survey of Employment Trends which indicate that the incidence of part-time employment is greater in small firms than in large ones, while the incidence of temporary employment is quite similar across firm sizes. 2/

What are the reasons behind the growing share of part-time employment? From the supply side, the female participation rate has been increasing since the mid-1970s. This is due both to an evolution in the perception of women’s role in society, as well as to economic incentives. 3/ However, many women continue to bear familial responsibilities, resulting in a desire to maintain the flexibility afforded by part-time employment. 4/ Moreover, part-time employees are not obliged to accede to company transfers. Public policies also provide incentives for part-time employment. For example, there is a relatively high threshold before secondary household earners’ incomes are subject to personal taxation. Moreover, the primary earner continues to receive a dependant deduction for tax purposes and the secondary worker remains eligible to receive a government pension. Also, most firms continue to provide health insurance coverage to the secondary worker under the primary worker’s plan, and to provide the primary worker with a family allowance.

There is also growing motivation for employers to hire part-time employees. As discussed more fully below, the increasing share of the middle-aged and elderly in the labor force, combined with a seniority-based pay structure, has increased worker’s compensation as a share of firms’ costs, and as a share of national income. In contrast, labor costs of part-time employees are lower than for regular “life-time” employees, although this must be balanced against potentially lower productivity due to less firm-specific human capital. Part-time employees generally receive fewer company-provided fringe benefits as well. Moreover, employers either are not required to pay or pay reduced contributions for unemployment insurance, pensions, and health insurance. 1/ In addition to the lower cost of part-time workers, Houseman and Osawa provide evidence that employers hire part-time and temporary workers to increase employment flexibility, and to provide a “buffer” for firms. 2/ It also is easier to dismiss part-time workers, or to simply not renew temporary and dispatched worker contracts. 3/ Thus, for both supply- and demand-side reasons, and because of company policies and the present legal structure, the share of part-time employment has risen over the past decade.

Another response of firms facing employment pressures that has been resorted to extensively in the recent downturn is the curtailment or halting of the hiring of new graduates. Firms have sharply restricted hiring recently, as seen in Chart II.6. The employment rate for junior college and university graduates has fallen by over 10 percentage points, to 70.5 percent in 1994. Similar, albeit somewhat smaller, reductions have also been seen among upper secondary and master’s degrees graduates. In 1995, a one-time survey by the Ministry of Labor on Employment Management found that about 70 percent of firms surveyed responded that they had not employed new school graduates. 4/

CHART II.6
CHART II.6

JAPAN: EMPLOYMENT RATES, 1955-94

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

In response to the weak labor market, students have extended their educations; the share of upper secondary students continuing their education has risen by 5.5 percentage points since 1990, to 36 percent, and 9.5 percent of junior college and university graduates continued their education in 1994, compared to 6.8 percent in 1990. Nevertheless, the unemployment rates for new graduates have increased steadily since 1990 for all but lower secondary and doctoral graduates. Again, females have borne a disproportionate share of the curtailment in hiring. The female unemployment rate of junior college and university graduates in 1994 was 16.3 percent, compared to 9.0 percent for males, 24.0 percent among female master’s degree graduates, compared to 6.3 percent for their male counterparts, and 33.0 percent among female doctoral graduates, compared to 20.4 percent for male doctorates. 1/

It should be noted that while the curtailment of new hiring helps reduce firms’ short-run overstaffing pressures, it is not without costs. For example, it increases the relative age of firms’ labor forces, further exacerbating wage bill pressures under seniority-based pay structures. To the extent that recent graduates become unemployed, their stocks of human capital begin to depreciate, although at present the seriousness of this problem is far less than seen in Western Europe. Therefore, this adjustment method is at best a short-run fix, with potentially serious consequences if relied on for too long.

There are two other methods for adjusting employment levels that have been used to some degree in the recent recession, although data on their use are generally less comprehensive. The first is the transfer of employees by firms with surplus labor (shukko). The Ministry of Labor Survey’s on Employment Management reports that during 1992-94, 60.8 percent of responding firms had enforced employment adjustment (including adjustment in hours worked). Of this total, 23.8 percent of responding firms had transferred employees in response to the need to adjust labor forces. 2/ These transfers generally took place from larger firms, and among manufacturing, wholesale trade, finance, and insurance firms. In general, larger firms have more integrated operations and more extensive relations among partner and affiliated firms. 3/ Of this total, most employees transferred came from the original enterprises’ manufacturing departments (42.7 percent), followed by 24.8 percent of management employees, while receiving firms’ manufacturing departments received 42.0 percent of transferred workers, followed by 33.4 percent of receiving firms’ sales and marketing departments.

The final and relatively drastic measure for labor force adjustment, given the prevalence of the lifetime employment system, is early retirement. The Ministry of Labor’s 1994 Industrial Labor Situation survey found that 11.7 percent of firms enforcing employment adjustment instituted voluntary retirement schemes. However, this practice clearly erodes the implicit guarantees provided under the lifetime employment system. 1/

5. Unemployment and labor market turnover

Aside from its lifetime employment system, the one aspect of Japan’s labor market that receives the most attention is it low unemployment rate. While many developed economies had similarly low rates prior to the first oil shock, Japan has seen only a small subsequent increase. 2/ Nevertheless, the unemployment rate has reached a post-war high of 3.4 percent.

There continues to be a popular misconception that the unemployment rate in Japan is in large part a “statistical quirk.” While unemployment rates are determined by the structure of the economy and the methodology of statistical classification and compilation, the latter in Japan are in accord with the suggested methodology of the International Labor Office (ILO). As a result, Japan’s actual and “internationally comparable” unemployment rates are the same. 3/ Of course, differences among countries in the detailed structure of otherwise internationally comparable questionnaires and classifications may remain, with the result that a country’s unemployment rate may vary when adjusted to another country’s methodology. Nevertheless, these changes for Japan have been found to be of minor importance (see Appendix).

In addition to having a low unemployment rate, the typical duration of unemployment spells in Japan are also relatively short. The top panel of Table II.2 includes data provided in Sorrentino (1995) on the share of the labor force that has been unemployed for 13 or more weeks. Only Sweden has had a smaller share of long-term unemployed, although a severe recession there has resulted in a sharp increase since 1991 in long-term employed. The bottom panel provides data from Sorrentino (1993) for shares of unemployment of various durations in 1989. Again, Sweden had somewhat lower long-term unemployment rates in this particular year, as did to a lesser degree, the United States. These data do indicate, however, that long-term unemployment is less severe in Japan than in most other major developed countries. Examining data in Table II.3 on the reasons for unemployment also suggest that unemployment generally does not occur in Japan as a result of losing one’s job. Only Italy and the Netherlands had higher levels of job security and protection, as reflected in the share of unemployed by job losers, than in Japan. However, Italy’s labor market posed significant difficulties for new entrants, and the share of re-entrants among total unemployed was the highest in the Netherlands.

Table II.2.

Japan: Unemployment Duration for Selected Countries, 1983-93

article image
Sources: Top panel, Sorrentino (1995); bottom panel, Sorrentino (1993).

Data for 1984-91 for West Germany; thereafter for unified Germany.

Table II.3.

Japan: Unemployment Rates by Former Status, 1983-93

article image
Source: Sorrentino (1995).

1984-93.

Not available separately.

1987-93.

1985-93.

1985-91.

1992-93.

1986-93

1983, 1985, 1987-91.

While these indicators point to relatively smooth and rapid transitions within Japan’s labor market, its unemployment rate began rising in late 1991 and now stands at a post-war high (see top panel of Chart II.7). As the lower panel shows, unemployment rates have increased almost identically for males and females, with little changes in their shares of total unemployment. 1/ Nevertheless, does the recent unprecedented rise in Japan’s unemployment rate indicate that its labor market has “broken down” in some sense, or is this the response one would expect, given the severity of the recent recession?

CHART II.7
CHART II.7

JAPAN: UNEMPLOYMENT RATES, 1970-95 1/

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

Shaded areas indicate recessions as defined by the EPA

This issue can be examined in two ways. First, one can analyze whether the unemployment rate is unusually high relative to the depth of the recession. Chart II.8 shows the relationship between the unemployment rate (plotted inversely) and the staff’s estimate of the output gap in Japan. The unemployment rate showed little variation prior to the first oil crisis, despite large swings in economic activity. In sharp contrast, the unemployment rate has tracked the business cycle quite closely since about 1980. 2/

CHART II.8
CHART II.8

JAPAN: OUTPUT GAP AND UNEMPLOYMENT RATE, 1960-95

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

In contrast, on the basis of developments in productivity, a number of observers have concluded that enterprises are “hoarding” large amounts of excess workers, and that the “true” unemployment rate is therefore significantly higher than official figures. As seen in the top left panel of Chart II.9, productivity in the manufacturing sector dropped sharply during 1990-93, and has stagnated for the overall economy as well. However, it may be more appropriate to examine changes in real unit labor costs (i.e., the labor share of output) to judge the degree to which there is excess employment, as this provides a better picture of the pressures on enterprise profitability. To the degree that the difference between the rate of increase in compensation relative to output prices declines, pressures from excess workers on firms’ profits are reduced. The lower left panel of Chart II.9 indicates that the trend increase in labor’s share of output has indeed moderated in the 1990s, although it continues to rise in manufacturing.

CHART II.9
CHART II.9

JAPAN: PRODUCTIVITY, PRODUCT WAGES AND EXCESS EMPLOYMENT 1980-95

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

The lower right panel presents estimates of the number of excess workers in the manufacturing sector and the overall economy, based upon deviations from trend real unit labor costs. 1/ The manufacturing sector has clearly displayed large swings in excess labor, with excess workers totaling 8.6 percent of its labor force in the fourth quarter of 1993 (equivalent to almost 1 million workers). But by the end of 1995 this sector has shed some 650,000 workers, while industrial output has recovered about one half of its lost ground, resulting in a sharp rebound in productivity, returning it to its trend level. As for the total economy, regular employment has been relatively stable since late 1992, while growth in labor’s product wage (defined as employee compensation per worker deflated by the GDP deflator) slowed. These results suggest that neither the total economy nor the manufacturing sector may have significant levels of excess employees. 2/

Second, the rising unemployment rate may reflect changes in the efficiency of the job matching process in Japan. This issue can be addressed by examining changes in the relationship between the unemployment rate and the rate of registered job vacancies. 3/ Over the course of the business cycle, one would expect a negative relationship between the two rates, and therefore a line with a negative slope when plotting the two series. As seen in Chart II.10, however, with the exception of the bubble period the vacancy rate has remained relatively stable since 1975, while the unemployment rate has steadily increased. This suggests that the “UV curve” has shifted outward, reflecting a deterioration in labor market efficiency.

Chart II.10
Chart II.10

Japan: Unemployment and Vacancy Rates, 1970-95

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

This deterioration could be the result of regional, age or gender mismatches. There is some evidence that regional and age mismatches have worsened. 1/ Therefore, while the increase in the unemployment rate appears to be largely as expected given the depth of the recent recession, and there no longer appears to be significant levels of excess workers, there is evidence that the efficiency in job matching has worsened somewhat.

6. Wage behavior

Japan is noted for its flexible wages, and it has been argued that this is in part the result of its shunto system of synchronized annual wage contracts. 2/ Table II.4 displays data on wage increases over the last decade. Nominal cash earnings increased fairly rapidly during the bubble period, as overtime hours increased and bonus payments surged. The share of bonus payments peaked in 1990 at over 40 percent of scheduled earnings. 3/

Table II.4.

Japan: Wage Developments, 1985-95

article image

Worsening economic conditions resulted in a sharp drop in hours worked, in overtime earnings, and, in 1993, in a decline in bonus payments. Declining inflation and the rising output gap led to smaller increases in the shunto, which at 2.8 percent in 1995, was the lowest nominal and real increase to date. Slight increases in overtime hours and in bonuses in 1995 offset the declining shunto, resulting in the same increase in nominal cash earnings as in 1994. 1/

7. Shifts in the lifetime employment system

As seen in the previous sections, the Japanese labor market faces significant pressures. In addition to the deep recession, structural shifts have accelerated due to the effects of a significant appreciation of the yen on competitiveness. 2/ Moreover, the rapid aging of the workforce, which, when combined with the seniority-base pay system, has exacerbated the cyclical increase in personnel expenditures as a share of total sales (Chart II.11). 3/

CHART II.11
CHART II.11

JAPAN: Personnel Expenses and Recurring Profits, 1960-95

(Share of sales; four-quarter moving average)

Citation: IMF Staff Country Reports 1996, 114; 10.5089/9781451820522.002.A002

One response to these pressures has been a flattening out of wage-tenure profiles, at least in the manufacturing sector, and especially for production workers (see Table II.1). This pattern, especially between 1990 and 1994, is somewhat surprising, as a number of observers found that profiles tended to flatten during boom periods, as demand increases for new graduates, thereby pushing up their relative wages. 4/ However, Mincer and Higuchi (1988) find support for the hypothesis that the steepness of wage-tenure profiles is related to productivity growth when comparing Japan and the United States. Japanese productivity growth has slowed sharply in the 1980s compared to previous decades, and even further in recent years. It remains an open question what the implications for wage-tenure profiles are from economic recovery, which should boost productivity somewhat, continued population aging, which increases total wage bills, and the reduction in the number of new school graduates. 1/

Anecdotal evidence suggests that the seniority-based pay system is also under assault. A Ministry of Labor survey conducted in December 1994 of 5,300 firms with 30 or more employees found that 4.3 percent of respondents had switched to merit-based pay systems (with 7.9 percent of firms with 1,000 or more employees having done so), while an additional 3.5 percent of all firms were considering doing so. A survey by the Economic Planning Agency in early 1996 found that while only 13.4 percent of responding firms had shifted to a merit-based system in the past three years, 44.6 percent planned to do so in the next three years, of which 98.6 percent stated that they intended to do so regardless of economic conditions. 2/ It also appears that a shift to a merit-based system is preferred by new graduates, with 72.4 percent favoring such a switch, according to a poll taken this spring by the Japan Productivity Center for Socio-Economic Development.

Mincer and Higuchi (1988) conclude that as much as two-thirds of the differential in turnover between Japan and the United States is explainable by the differences in the steepness of wage profiles. 3/ Therefore, a flattening of wage profiles and/or a shifting from the seniority-based wage system would lessen the incentives for individuals to remain with the same firm for their working lives, as the costs of switching jobs would be reduced. A recent government report stated that a record 5.2 million people wanted to switch jobs in 1995, with 1.95 million actually looking for new jobs. The previously cited survey of new graduates also found that 64.9 percent of respondents would switch jobs to get better conditions, while 31.4 percent replied they would jump immediately to a different job if offered better conditions. More telling for the future of the lifetime employment system, 51.8 percent of new graduates said that the appropriate period of time to work for one company was only 2.5 years.

8. Government labor programs and restrictions

The government has undertaken efforts to improve labor market conditions and efficiency in recent years. While not comprehensive, the following include the more important, as well as some recently adopted, programs.

In 1975, the government replaced the Unemployment Insurance Act with the Employment Insurance Law. While continuing to provide unemployment benefits to those eligible, the law also created a system of employment adjustment subsidies for “designated” industries experiencing specified reductions in output or hours worked. These subsidies are paid to the employers in the designated industries to provide vocational training for workers to be seconded to other firms, to employers in other nonaffected industries to accept workers from designated industries and to train them, and to employers in designated industries who invest in new businesses to maintain employment. In total, 94 industries were designated as eligible for these subsidies at end-1995, and 1.57 million workers were covered. The law was amended in late 1995 to include a program to aid small business (through wage subsidies) attempting to create new ventures.

In 1987, in response to the labor market pressures resulting from the yen’s sharp appreciation, a system of direct tuition subsidies to middle aged individuals who undertook training was established. Since 1990, the age for eligibility has been successively lowered and the length of eligible training shortened. 1/

In response to the disproportionate burden of labor force adjustment falling on new graduates, the government has asked the Japanese Federation of Employers (Nikkeiren) to boost their efforts to promote hiring, and is attempting to actively provide information concerning job opportunities and vocational guidance. Student employment and vocational counseling centers, which have existed in major cities, have been expanded to 41 prefectures. Unified lists of job offers for new graduates were compiled in July and November 1995, and March 1996, and were supplied to universities and student employment centers. The centers also hosted meetings bringing together prospective graduates and employers throughout country. The Ministry of Labor intends to establish by the end of the decade “work experience plazas” for non-university graduates, intended to provide job information, and to enhance youth understanding about vocational possibilities.

Despite these efforts, a number of impediments exist that inhibit labor mobility. First, private employment placement services are restricted to jobs approved by the Ministry of Labor that require special skills, and that can not be covered by public employment centers. While 29 types of jobs are presently designated, with some covering broad categories (e.g., engineer), restrictions on private placement services clearly reduces information and job mobility. Second, temporary employment agencies have been limited since 1986 to be engaged in only 16 categories of jobs (including secretaries, interpreters, receptionists and computer operators). While the government intends to broaden the scope of these activities by adding 12 additional categories next year, the scale of temporary employment in Japan lags far behind that in the United States, or even in some Western European nations. Third, the Labor Standards Law includes restrictions concerning the number of overtime hours that women can work, as well as a ban on night work. Again, a large number of exceptions apply, including many managers and professionals, but these restrictions limit women’s participation in many production jobs. Fourth, retiring workers in Japan generally receive from their employers a lump-sum payment, although an increasing number receive pensions. Regardless of the form of payment, a special deduction is available for retirement income for tax purposes. However the amount deductible is a fixed multiple of the number of years of service until 20 years, after which the multiple shifts to a higher level. Therefore, the tax treatment of retirement income penalizes workers with less than twenty years service, inhibiting labor mobility.

9. Prospects and conclusions

There are increasing signs that a steady, albeit mild, economic recovery is underway in Japan, which is beginning to be reflected in labor market conditions. While employment in agriculture and among nonagricultural self-employed and family workers continues to decline, the number of corporate employees has risen since last autumn. However, it is thought that many of these jobs are part-time, reflecting a continuing hesitation of employers to increase their permanent labor forces. 1/ Scheduled hours worked have also increased by about 1/2 percent per quarter in the end of 1995 and first quarter of 1996, while overtime hours increased by 4 percent in the first quarter. The 1996 shunto wage agreement of 2.8 percent was the first time that the rate of increase had not fallen in 5 years. And while the labor market continues to be especially difficult for new graduates, many firms in various sectors of the economy (including major manufacturers) have announced their intentions to increase the hiring of new graduates in 1997.

While the recovery will help ease labor market pressures, adjustments will continue to occur for structural reasons. Although a shifting of production from domestic sources to overseas operations has been underway for some time, and has been accelerated by the yen’s rise in recent years, the share of overseas production in total production for Japanese industry remains lower than comparable figures for the United States and Germany. While various determinants influence these shares, and they are not necessarily expected to converge, continued strong foreign direct investment by Japanese firms, especially in Southeast Asia, as well as the coming online of recent investments, suggest that the shifting of jobs is not complete.

Rapid population aging will also lead to labor market pressures. The dependency ratio, defined as the ratio of population less than 15 years and more than 64 years to those 15-64 years of age, is projected to increase from about 30 percent in 1995 to over 40 percent by 2020, before rising further to over 44 percent in 2045. This increasing ratio will put downward pressure on living standards unless one or more of the following things transpires: unemployment is reduced; participation rates are increased; labor is reallocated to higher productivity sectors; productivity is increased in general. Clearly there are limits to lowering the unemployment rate and raising the retirement age. Therefore, sustaining and further increasing living standards requires boosting productivity. 1/

Japan has had an enviable record in the smooth operating of its labor force and high level of labor mobility. Structural reforms and labor market aging should work to flatten tenure-earnings profiles, and shifts toward merit-based pay systems could also increase job mobility. In this regard, it is important that impediments to mobility be minimized. Among those that the government could address is a lifting of limits on private employment and temporary placement activities (or at the least a shifting from a system of limited approved areas to one of unlimited access with limited exceptions), an a repeal of overtime limits for female workers.

APPENDIX: Japan’s Unemployment Rate

Japan’s low unemployment rate has often been dismissed as a statistical artifact, with references to an unusually liberal definition of employment, and restrictive definition of unemployment. Although potentially significant differences in unemployment rates in Japan and other developed economies do arise from differences in economic structure and behavior, differences arising from methodological approaches in defining and measuring unemployment rates are of secondary importance.

I to (1984) compared the United States and Japanese definitions of unemployment, finding that these differences have only a minor effect on measured unemployment rates. 1/ Instead, he concluded that at least half of the difference in measured unemployment rates was attributed to differences in economic structure. Most importantly, he found that 20-25 percent of the difference was due to low participation rates of teenagers in Japan relative to the United States. Differences in layoff practices and the relative shares of the agricultural sector (which are largely self- or family-employed and therefore free from measured unemployment), also accounted for a significant share of the difference. Ito (1992) again concluded that statistical definitions in the United States and Japan did not explain the gap in measured unemployment rates.

Hashimoto (1993) examined the role that kyogyosha, or those not at work, played in Japan’s low measured unemployment rate. These individuals are counted as employed, as they are kept on firms’ payrolls and receive wages or salaries. A portion of the workers that are temporarily not at work (ichiji kyogyosha, which corresponds to temporary layoffs in the United States) are also classified as employed in Japan, while they would be unemployed in the United States. He found some evidence that the share of kyogyosha in the population varies inversely with the business cycle. Although data limitations do not allow for a separation of temporarily and permanently not at work, the total number of kyogyosha could be used as an upper bound for the former, which would allow for an adjustment in Japan’s unemployment rate to compare it with that in the United States. Hashimoto concluded that such an adjustment would reduce the gap between the two official rates in 1988 by 43 percent (during which the United States, unemployment rate was 5.5 percent, while Japan’s official rate of 2.5 percent, and its adjusted rate was 3.8 percent). However, given the Japanese preference for adjusting hours worked rather than using temporary layoffs, this estimate must be viewed cautiously.

The most comprehensive work undertaken in comparing unemployment rates among developed countries using comparable methodologies has been by conducted by Sorrentino (1984, 1987, 1989, 1993, 1995). In Sorrentino (1984), she adjusted Japanese data using special March surveys to make them comparable with United States classifications, and found that this resulted in an increase in the Japanese unemployment rate by at most 0.4 percentage points, with the male unemployment rates revised downward and female rates revised upward. She also compared broader definitions of unemployment in Japan and the United States. 1/ When the conventional concept of unemployment was augmented to include most part-time workers, or include the latter as well as discouraged workers, she found that the broader measures were increased to a greater degree in Japan than in the United States (i.e., that part-time workers for economic reasons and discouraged workers were proportionately larger in Japan than in the United States). These calculations were updated in Sorrentino (1987, 1989). First, she found that the use of annual special surveys enabling the adjustment of Japanese data for these later years now indicated that the Japanese unemployment rate was slightly lower than the United States rate when placed on a comparable basis, although the differences remained minor. Second, the adjusted female unemployment rates remained higher than their unadjusted counterparts, while adjusted male rates were again lower. Third, part-time jobseekers for economic reasons and discouraged workers continued to be disproportionately present in Japan. In fact, when discouraged workers were included, the Japanese unemployment rate exceeded its United States counterpart during 1985-88. 1/ 2/

In Sorrentino (1993), international comparisons for 1989 were expanded to include a number of other OECD countries, as well as the Japan and the United States. 3/ While Sweden and Japan had the lowest standard unemployment rates, the increase in the rate when including part-time workers for economic reasons and discouraged workers was the highest in Japan, especially for women. These patterns were confirmed in Sorrentino (1995), which compiled data for the period 1983-93, although caution was again advised in interpreting the latter results, given international differences in concepts of discouraged workers. 4/

Finally, the OECD (1995) examined supplementary measures of labor market slack. It found that involuntary part-time work was a more important element of labor market slack in OECD economies with the exception of Japan, where discouraged workers were more important. However, it questioned whether all discouraged workers should be considered as additional labor market slack, finding that, among other reasons, most discouraged workers had been jobless for long periods of time, and that in European countries for which data were available, their numbers exhibited little cyclical variation. Nevertheless, in comparison to many other countries, Japan’s discouraged worker rate was highly correlated with its unemployment rate. Thus, it would appear that this is an aspect of Japan’s labor market that require’s further study. 1/

References

  • Abraham, Katherine, and Susan Houseman,Job Security and Work Force Adjustment: How Different are U.S. and Japanese Practices?,Journal of the Japanese and International Economies. 3, pp. 50021, 1989.

    • Search Google Scholar
    • Export Citation
  • Bank of Japan, “Recent Changes in Japan’s Labor Market and Their Impact -Synopsis,” Bank of Japan Special Paper No. 202, May 1991.

    • Search Google Scholar
    • Export Citation
  • Bank of Japan, “The Japanese Employment System,” Quarterly Bulletin. May 1994.

  • Bank of Japan, Comparative Economic and Financial Statistics: Japan and other Major Countries. 1995. Tokyo, 1995a.

  • Bank of Japan, “Corporate Profits and Restructuring Measures during the Recent Economic Downturn,” Quarterly Bulletin. February 1995b.

    • Search Google Scholar
    • Export Citation
  • Baumgartner, Ulrich, and Guy Meredith (eds.), “Saving Behavior and the Asset Price ‘Bubble’ in Japan: Analytical Studies,” IMF Occasional Paper #124, April 1995.

    • Search Google Scholar
    • Export Citation
  • Blumenthal, Tuvia,Labor Adjustment Policy in Japan,Hitosubashi Journal of Economics, 35 (1994), 95108.

  • Bregger, John, and Steven Haugen,BLS introduces new range of alternative unemployment measures,Monthly Labor Review. October 1995, pp. 1926.

    • Search Google Scholar
    • Export Citation
  • Brunello, G.,Reverse seniority rules and the responsiveness of wages and employment to external shocks: A note on the Japanese experience,” The Economics Studies Quarterly, 39(3), pp. 208215, 1988.

    • Search Google Scholar
    • Export Citation
  • Brunello, G., Kenn Ariga, Yohihiko Nishiyama and Yasushi Ohkusa,Recent Changes in the Internal Structure of Wages and Employment in Japan,Journal of the Japanese and International Economies, 9, pp. 105129, 1995.

    • Search Google Scholar
    • Export Citation
  • Clark, Robert, and Naohiro Ogawa,Employment Tenure and Earnings Profiles in Japan and the United States: Comment,American Economic Review. Vol. 82, No. 1, March 1992, pp. 336345.

    • Search Google Scholar
    • Export Citation
  • De Masi, Paula, and S.G.B. Henry,Aspects of the Swiss Labor Market,IMF Working Paper 96/3, January 1996.

  • Economic Planning Agency, Social and Economic Plan for Structural Reforms: Towards a Vital Economy and Secure Life (Summary). Tokyo, November 29, 1995.

    • Search Google Scholar
    • Export Citation
  • Economic Planning Agency, 1996 Regional Economic Report.

  • Feldman, Robert,The Golden Goose and the Silver Fox: Data [Draft],” Solomon Brothers (Tokyo), May 14, 1996;

  • Freeman, Richard, and Martin Weitzman,Bonuses and Employment in Japan,Journal of the Japanese and International Economies. 1, pp. 168194, 1987.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, Masanori,Aspects of Labor Market Adjustment in Japan,Journal of Labor Economics. Vol. 11, No. 1, Part 1, January 1993.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, Masanori, and John Raisian,Employment Tenure and Earnings Profiles in Japan and the United States,American Economic Review. September 1985, pp. 72136.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, Masanori, and John Raisian,Investments in Employer-Employee Attachments by Japanese and U.S. Workers in Firms of Varying Size,Journal of the Japanese and International Economies. 3, pp. 3148, 1989.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, Masanori, and John Raisian,Employment Tenure and Earnings Profiles in Japan and the United States: Reply,American Economic Review. Vol. 82, No. 1, March 1992, pp. 346354.

    • Search Google Scholar
    • Export Citation
  • Higuchi, Yoshio,Japan’s Changing Wage Structure: The Impact of Internal Factors and International Competition,Journal of the Japanese and International Economies. 3, pp. 48099, 1989.

    • Search Google Scholar
    • Export Citation
  • Higuchi, Yoshio,Labor Turnover Behavior: Japan Versus the West,Japanese Economic Studies. Fall 1993, pp. 6188.

  • Hill, M. Anne,Female Labor Supply in Japan: Implications of the Informal Sector for Labor Force Participation and Hours of Work,The Journal of Human Resources, XXIV, 1, pp. 143161.

    • Search Google Scholar
    • Export Citation
  • Houseman, Susan and Machiko Osawa,Part-time and temporary employment in Japan,Monthly Labor Review. October 1995, pp. 1018.

  • Ito, Akihiko, Letter to Heads of National Statistical Agencies, Statistics Bureau of Japan, 1995.

  • Ito, Takatoshi,Why is the Unemployment Rate so much lower in Japan than in the U.S.?,Center for Economic Research Discussion Paper. No. 198, University of Minnesota, January 1984.

    • Search Google Scholar
    • Export Citation
  • Ito, Takatoshi, The Japanese Economy. The MIT Press, Cambridge, Massachusetts, 1992.

  • Jackman, R., R. Layard and C. Pissarides,On Vacancies,Oxford Bulletin of Economics and Statistics. 51, 4, 1989, pp. 37794.

  • Kurosaka, Yoshio,The Japanese Economy and the Labor Market,” Japanese Economic Studies. Summer 1989, pp. 340.

  • Management and Coordination Agency, Annual Report on the Labor Force Survey. 1995.

  • Matsuoka, Mikihiro, and Brian Rose, The DIR Guide to Japanese Economic Statistics, Oxford University Press, Oxford, 1994.

  • Mincer, Jacob, and Yoshio Higuchi,Wage Structures and Labor Turnover in the United States and Japan,Journal of the Japanese and International Economies. 2, pp. 97133, 1988.

    • Search Google Scholar
    • Export Citation
  • Ministry of Labor, Yearbook of Labor Statistics. Ministry of Labor, Tokyo, various years.

  • OECD, Employment Outlook. Paris, 1995.

  • Ohashi, Isao,On the Determinants of Bonuses and Basic Wages in Large Japanese Firms,Journal of the Japanese and International Economies. 3, pp. 45179, 1989.

    • Search Google Scholar
    • Export Citation
  • Ohkusa, Yasushi, and Souichi Ohta,An Empirical Study of the Wage-Tenure Profile in Japanese Manufacturing,Journal of the Japanese and International Economies. 8, pp. 173203, 1994.

    • Search Google Scholar
    • Export Citation
  • Ono, Akira,Labor Cost in an Aging Economy,” Japanese Economic Studies. Winter 1989/90, pp. 3057.

  • Sakurai, Kojiro, and Toshiaki Tachibanaki,Estimation of mis-match and U-V analysis in Japan,” Japan and the World Economy. 4, (1992), pp. 319332.

    • Search Google Scholar
    • Export Citation
  • Sato, Hiroki,Employment Adjustment of Middle-aged and Older White-collar workers,Japan Labor Review, February 1, 1994, pp. 58.

    • Search Google Scholar
    • Export Citation
  • Schregle, Johannes,Dismissal protection in Japan,International Labor Review. Vol. 132, No. 4, 1993.

  • Sorrentino, Constance, “Japan’s low unemployment: an in-depth analysis,Monthly Labor Review. March 1984, pp. 1827.

  • Sorrentino, Constance, “Japanese unemployment: BLS updates its analysis,Monthly Labor Review. June 1987, pp. 4753.

  • Sorrentino, Constance, “Adjusted Japanese unemployment rate remains below 3 percent in 1987-88,Monthly Labor Review. June 1989, pp. 3638.

    • Search Google Scholar
    • Export Citation
  • Sorrentino, Constance, “International comparisons of unemployment indicators,Monthly Labor Review. March 1993, pp. 324.

  • Sorrentino, Constance, “International unemployment indicators, 1983-93,Monthly Labor Review. August 1995, pp. 3150.

  • Tachibanaki, Toshiaki, and Kojiro Sakurai,Labor supply and unemployment in Japan,European Economic Review. 35, 1991, pp. 15751587.

    • Search Google Scholar
    • Export Citation
  • Taylor, John,Differences in Economic Fluctuations in Japan and the United States: The Role of Nominal Rigidities,Journal of the Japanese and International Economies. 3, pp. 127144, 1989.

    • Search Google Scholar
    • Export Citation
1/

This chapter was prepared by Mark S. Lutz.

2/

See Hashimoto and Raisian (1985), Bank of Japan (1994) and Ito (1992) for good overviews of Japan’s labor market structure and behavior, kurosaka (1989) provides a history of the Japanese labor market in the postwar period.

1/

By age 64, a typical Japanese male has held 4.9 jobs, while an American worker has held 10.9.

2/

Median tenure for firms with 1-9 workers in Japan was 8 years, compared to 2 years for firms with 1-25 workers in the United States, while for firms with more than 1,000 male employees it was 12 years in Japan and 7 years in the United States.

3/

It must be kept in mind, however, that these patterns may be influenced by differences in labor force composition (including age, educational background and industrial mix).

4/

These finding were supported by Clark and Ogawa’s (1992) and Hashimoto and Raisian’s (1992) examination of this issue using data from later time periods.

5/

These figures exclude nonself-employed and family workers.

1/

Hashimoto (1993) provides evidence of the emulation by small firms of large firms’ lifetime employment practices.

2/

What has been noticeably left unanswered concerns the participation of female workers in the lifetime employment system. As discussed more fully below, although the proportion of female employment in regular (i.e., not part-time, temporary or daily) nonagricultural employment has risen significantly in the past decade, women continue to be far less represented than men in the lifetime employment system, and shoulder a disproportionate share of the burden in employment adjustment.

3/

See Baumgartner and Meredith (1995) for a discussion of the “bubble” period.

4/

The Ministry of Health and Welfare projects that Japan’s total population will peak at about 130 million by about 2010, declining thereafter. The working age population, defined as those between 15 and 64 years of age, peaked in absolute terms in 1994 and is expected to decline as a share of the population from 69.7 percent in 1991 to less than 60 percent by 2020.

5/

See Bank of Japan (1994).

1/

See Bank of Japan (1991) and Tachibanaki and Sakurai (1991) for brief discussions of developments in female participation rates.

2/

One possible explanation for this pattern could be that demand shocks were anticipated to be of a more temporary nature in Japan than in the United States Hashimoto tests for this by comparing coefficient values in equations using estimates of anticipated and unanticipated shifts in demand. The coefficients for unanticipated demand indicate greater employment response in the United States, and greater inventory response in Japan, with little difference in adjustment in hours worked. While many of the coefficients for employment and hours worked in response to anticipated changes in current and future output were statistically insignificant, Japanese firms continued to significantly adjust inventories.

1/

Among their findings, in response to a 1 percent decline in production, employment in the United States declined by 0.3 percent within one month and by 0.8 percent within 12 months, compared to a 0.02 percent decline within one month and 0.2 percent within 12 months in Japan. While similar differences were found in total production hours worked, average production hours declined in both countries by about 0.2 percent within one month and by only 0.1 percent within 12 months.

2/

Abraham and Houseman (1989), page 514.

3/

Nevertheless, the Bank of Japan (1995a) reports that in 1993, the depth of Japan’s recession, total annual working hours in manufacturing averaged 2,014 in Japan, compared to 2,153 in the United States, 1,857 in the United Kingdom, 1,684 in France, and 1,542 in Germany.

1/

Hashimoto (1993) notes that overtime hours account for a greater share of total hours worked in Japanese manufacturing firms than in the United States, which may result from a smaller overtime premium in Japan (time and a quarter, compared to time and a half in the United States), as well as from the influence of the firm-specific human capital.

2/

The equations regressed quarterly changes in the logs of employment and hours worked on contemporaneous through four lags in the change in the logs of manufacturing industrial production for the manufacturing sector and the same pattern of changes in the logs of real GDP for the overall economy over the 1973-90 period. Time trends were included in both equations. Predictions were then made for the 1991-95 period. This method of estimating employment and hours worked elasticities has also been used by Abraham and Houseman (1989) and Bank of Japan (1994). Hashimoto estimated a similar model using annual data, while Houseman and Osawa (1995) looked at adjustment of part-time workers using quarterly data. The chart contains four-quarter moving averages of the prediction errors to smooth the series and accentuate persistent errors.

1/

The Employment Status Survey is conducted every five years by the Management and Coordination Agency. Part-time workers are separately enumerated, as are arubaito, defined as side jobs taken by someone who is in school or who has regular employment elsewhere. The feature distinguishing part-time workers from other regular employees is not necessarily the number of hours worked, but the set of personnel practices that are applied to them. As discussed below, for example, part-time workers do not participate in the lifetime employment and seniority-based wage system, nor are they generally promoted. The annual Labor Force Survey does not separately track part-time workers, but annual data on the number of nonagricultural employees categorized by number of hours worked are available. Given that some part-time workers in fact work the same number of hours as regular full-time workers, the annual data most likely understate the share of part-time employees in the labor force. The definitions of temporary and daily workers are the same in the two data sources.

2/

See Houseman and Osawa (1995) on differences in definitions and coverage in the Ministry of Labor Survey of Employment Trends and other data sources.

3/

See Hill (1988) for estimates of female labor supply in the Tokyo area.

4/

Of the 95 percentage points accounted for by female part-time employees, women aged 30-39 accounted for 20 percentage points, while those aged 40-49 accounted for 37 percentage points, and aged 50-59 accounted for 22 percentage points. Therefore, women generally wait until their children reach school age before re-entering the labor force.

1/

See Houseman and Osawa (1995) for details of reduced contributions for part-time employees.

2/

They estimated quarterly employment elasticities for regular and temporary manufacturing employees over four quarters, finding significantly higher elasticities for the latter group in all circumstances except the one-month elasticity for men. Hashimoto (1993) provides similar evidence on the differences in employment elasticities for regular versus temporary and day laborers, and for females and males using annual data.

1/

One must, of course, keep in mind that differences in subject specialization influences these differences.

2/

This compares to 23.7 percent of responding firms that reduced or suspended the recruitment of employees in mid-careers, and 20.6 percent that reduced or suspended the recruitment of new school graduates. See Ministry of Labor (1994).

3/

See Blumenthal (1994) for a discussion of inter- and intra-firm employment adjustment. In order to enhance smaller firms’ abilities to engage in inter-firm transfers, and spurred by the yen’s appreciation in the mid-1980s, the Industrial Employment Stabilization Center was established in 1987. It conducts biannual surveys of companies with redundant and insufficient workers, disseminating the results, and thereby facilitating transfer agreements.

1/

See Sato (1994) for a discussion of increasing pressures from excess middle-aged and older workers.

2/

Until recently, Switzerland was able to boast an unemployment rate that was lower than Japan’s. While prior to the 1980s this was in part the result of limited statistical coverage of the unemployed, it has also reflected variations in female participation rates, as in Japan, and In the cyclical supply of foreign labor. See De Masi and Henry (1996).

3/

And these are the data included in the OECD’s standardized unemployment rates.

1/

Since 1991, unemployment rates have risen for all age groups, with the largest increase occurring among workers aged 15-24 years (rising to 6.1 percent in 1996). Unemployment among workers aged 25-34 years and aged 55-64 years have also risen markedly, to 3.7 and 3.8 percent, respectively, in 1995.

2/

Over the 1985-95 period, a simple regression of the unemployment rate on the output gap, lagged two periods, had a adjusted R squared of 90, with a 1 percentage point increase in the output gap resulting in a 0.14 percentage point decrease in the unemployment rate.

1/

The data and estimates in Chart II.9 are for regular employees (excluding temporary and part-time workers) with the trend real unit labor costs derived using a Hodrick-Prescott filter.

2/

While these estimates are somewhat sensitive to the time period examined and method for trend estimation, alternate estimates using total, rather than regular, employment, and adjusting for hours worked provide similar results.

3/

See Jackman, Layard and Pissarides (1989) for a discussion of unemployment/vacancy rate issues in the United Kingdom. For Japan, data on job vacancies are collected via some 600 public employment agency offices, and are compiled by the Ministry of Labor. However, these statistics are thought to cover only about 20 percent of job searches (see Matsuoka and Rose, 1994).

1/

Sakurai and Tachibanaki (1992) provide evidence suggesting that regional and age mismatches account for 17-20 percent of total unemployment over the 1974-86 period. More recently, the Economic Planning Agency’s annual regional economic report concluded that recent industrial hollowing out has been felt most heavily in the regions including Tokyo, Osaka and Nagoya, where subcontractors of large manufacturers tend to be located. Moreover, the report notes that there has been an increase in workers moving to provincial areas from large cities. As to age mismatches, unemployment rates have risen the most for youths aged 15-24, and those aged 55-64. Higuchi (1989) discusses the longer duration of unemployment for senior workers due to a scarcity of reemployment opportunities.

2/

See Kurosaka (1989) for a history of the shunto system, and Taylor (1989) for a comparison of wage flexibility in Japan and the United States. Brunello (1988) and Higuchi (1989) argue that higher labor turnover costs from the seniority-based pay system, especially for older workers, induces greater wage flexibility.

3/

There has been a sizable debate over the role played by Japan’s system of bonus payments. For example, Freeman and Weitzman (1987) have found bonuses to be more cyclical than base wages, arguing that they contain a significant profit-sharing component, which has played a modest role in stabilizing unemployment at relatively low levels. They found that bonuses are also more sensitive to a company’s or industry’s particular circumstances, while base wages are primarily influenced by the overall economy’s performance. Ohashi (1989) argues that work intensity plays a greater role than profits in influencing bonus patterns, and therefore views them in an efficiency wage context. Taylor (1989) discounts the role bonuses play in Japan’s wage flexibility, emphasizing instead the benefits of synchronized annual wage setting.

1/

The recently completed 1996 shunto was also 2.8 percent. These increases appear to be influenced by both inflation and the output gap. A regression of the annual shunto on current and lagged inflation, and on the lagged output gap, with instruments used for current inflation, has an adjusted R squared of 0.92, and a Durbin-Watson statistic of 1.89.

2/

For example, in a Ministry of Labor survey of 4,500 manufacturers with 100 or more employees conducted in early 1996, 25.3 percent of respondents stated that they had reduced their workforces due to increased imports and an expansion in overseas production.

3/

Ono (1989/90) estimates the between 1961 and 1985, the average age of manufacturing workers rose by about nine years, and the proportion of workers over fifty years of age more than doubled, from 7.1 percent to 17.7 percent. Data from the 1995 Labor Force Survey indicate that the share of total employees in the economy over fifty rose from 28.4 percent in 1985, which is much higher than Ono’s figure, to 33.0 percent in 1995. While the absolute levels in the two sources differ, the sharp increase in the share of older workers is clearly apparent. Higuchi (1993) argues that it was Japan’s relatively voung labor force that enabled the lifetime employment system to operate.

4/

See Ohkusa and Ohta (1994) and Higuchi (1989). Recent wage increases for new college and university graduates and for existing workers contained in the shunto also support the flattening of profiles during boom periods, as the former were almost equal to the latter during the bubble period, but have been about 1/4 of the latter during 1992-95.

1/

The Ministry of Health and Welfare projects that the population in their 20s will decline from almost 19 million in 1995 to 14 million in 2010.

2/

Independent announcements have been made by a number of major steel and auto producers, as well as other manufacturers and financial institutions, that they have either adopted or intend to adopt by 1997 merit-based pay for at least a portion of their workforce (generally managers initially).

3/

They also attempt to standardize for the cultural background of workers by examining a sample of Japanese plants in the United States that employ American workers and use Japanese labor policies in recruitment and training. They find that steeper wage-tenure slopes and lower turnover rates place this sample closer to Japanese patterns.

1/

Also in 1987, the Industrial Employment Stabilization Center was founded as a non-profit organization, financed by member trade organizations and supervised by the Ministry of Finance, whose purpose is to facilitate labor transfers using a database collected through twice-yearly surveys of firms redundant and insufficient workers. The center has offices in every prefecture as well the three largest cities in Japan. The scale of the Center’s activities remain small however; in 1995 it handled only 1,165 cases involving 2,709 individuals.

1/

Increases in nonagricultural employment have centered in construction and services, sectors with relatively high shares of temporary and day workers, while manufacturing employment has continued to decline. The Association of Job Journals reports that vacancy advertisements rose 28 percent (year/year) in April 1996, although most of the jobs were part-time. Moreover, although restrictions on temporary placements limit their overall contribution to employment, the Ministry of Labor reported a 70 percent increase in the number of newly opened temporary agencies in the first five months of 1996, compared to the same period of 1995. Finally, the Temporary Work Service Association reported that the number of temporary placements rose by 24 percent in the latter half of 1995 over the same period in the previous year.

1/

It is this stark prospect that motivates the government’s focus on improving productivity in its latest five year plan (EPA, 1995). The plan includes two scenarios, based on whether deregulation and structural reforms are adopted or not, with annual growth projected to average about 3 percent over 1996-2000 with reforms, and only 1 3/4 percent otherwise. See also Feldman (1996) for an analytical examination of implications of demographic projections and productivity gains for living standards in the coming decades.

1/

Two factors contributing to a downward bias in the Japanese unemployment rate were the inclusion of the Self Defense Force as employed in Japan (although the United States did publish an unemployment rate including the armed forces during the 1983-93 period; see Bregger and Haugen (1995)), and the classification of unpaid family workers who worked fewer than 15 hours as employed in Japan (while they were counted as “not in the labor force” in the United States). Ito found, however, that these adjustments raised Japanese unemployment rates by only 0.1-0.2 percentage points.

Ito also compared the definitions of unemployed in the two countries. He found that the definition on unemployed in Japan was stricter by requiring jobseeking activity during the survey week, rather than in the past month as in the United States, thereby biasing down the unemployment rate. On the other hand, the kinds of activities required for jobseeking were broader in Japan, with a resulting upward bias. The overall effect of these biases on the Japanese unemployment rate was therefore an empirical question, which Ito concluded was of relatively minor importance.

1/

The United States Bureau of Labor Statistics compiled data over 1976-93 for alternate measures of labor force underutilization with varying degrees of strictness, labeled U1-U7. These ranged from a long-duration unemployment rate (Ul), generally the most restrictive measure, through the conventional unemployment rate (U5), to the conventional rate including most persons working part-time for economic reasons (U6), and also including discouraged workers (U7). See Sorrentino (1993) for details of the various concepts.

1/

It should be noted that in these two articles, the BLS used a range for discouraged workers in Japan to compare to their United States counterparts, due to differences in survey questions in the two countries, as well as the elusive nature of measuring discouraged worker attachments to the labor force. See the appendix to Sorrentino (1987) for a discussion of this issue.

2/

Tachibanaki and Sakurai (1991) estimate labor supply functions in Japan, focusing on the discouraged worker effect. They find that the unemployment rate would be 46 percent more variable over the 1963-80 period if these effect were absent.

3/

For technical reasons, data for Japan were from 1990.

4/

Calculations of U7, the unemployment rate which includes discouraged workers, was modified for Japan, compared to those presented in Sorrentino (1993), resulting in somewhat lower estimates. See the appendix to Sorrentino (1995).

In addition, internationally comparable data were not compiled by the BLS after 1993 because a redesign of the U.S. Current Population Survey in 1994 resulted in revised questions regarding employment and unemployment activities which did not allow for the earlier adjustments of international data. Most importantly, the BLS’s new alternate measures of unemployment and other forms of labor force underutilization result in sharply lower estimates of persons classified as employed part-time for economic reasons (as respondents were now explicitly asked about their desire and availability for full-time work), and a considerable tightening of the requirements for discouraged worker status (resulting in a 50 percent reduction in this group). See Bregger and Haugen (1995) for a discussion of BLS’s new range of alternative unemployment measures, and how they differ from the old alternate unemployment measures.

1/

In private correspondence with Directors of national statistical agencies, Akihiko Ito, the Director General of Japan’s Statistics Bureau, addressed the issue of discouraged workers in Japan. Under the BLS’s old U1-U7 system, the rate of discouraged workers in Japan was much higher than that in other major developed countries, totaling 2.85 million workers in February 1995. However, the number of discouraged workers varies substantially with the definition for discouragement. Under the new BLS definition, which requires discouraged workers to be able to take a job immediately (while no time dimension had been included before) and to have looked for a job during the past year (which was not previously required), the number of discouraged workers fell to only 330,000, a reduction of a 88 percent, compared to a 50 percent reduction in the United States data. As a result, the United States’ unemployment rate was 5.9 percent in February 1995 (the new U3 rate), while the Japanese rate on a comparable basis was 2.8 percent (with the official rate of 2.9 percent). Were the new definition of discouraged workers included (the new U4 measure), the United States rate would have risen to 6.2 percent, while the Japanese rate would have risen to 3.9 percent. Thus, although the modified unemployment rate rises by a greater percentage in Japan, as it did under the old definition of discouraged workers, indicating that discouraged workers remain a proportionately larger problem in Japan, the level of the unemployment rate remains below its United States counterpart, in contrast to calculations in Sorrentino (1989).

  • Collapse
  • Expand
Japan: Selected Issues
Author:
International Monetary Fund
  • CHART II.1

    JAPAN: WORKING AGE POPULATION, LABOR FORCE AND EMPLOYMENT GROWTH, 1960-95

  • CHART II.2

    JAPAN: LABOR FORCE PARTICIPATION RATES, 1960-95

  • CHART II.3

    JAPAN: COMPARISON OF FOUR CYCLICAL DOWNTURNS 1/

    Index (trough of cycle=100)

  • CHART II.4

    JAPAN: HOURS WORKED, 1970-95 1/

  • CHART II.5

    JAPAN: EMPLOYMENT AND HOURS WORKED PREDICTION ERRORS, 1974-95

    (Four-quarter moving average)

  • CHART II.6

    JAPAN: EMPLOYMENT RATES, 1955-94

  • CHART II.7

    JAPAN: UNEMPLOYMENT RATES, 1970-95 1/

  • CHART II.8

    JAPAN: OUTPUT GAP AND UNEMPLOYMENT RATE, 1960-95

  • CHART II.9

    JAPAN: PRODUCTIVITY, PRODUCT WAGES AND EXCESS EMPLOYMENT 1980-95

  • Chart II.10

    Japan: Unemployment and Vacancy Rates, 1970-95

  • CHART II.11

    JAPAN: Personnel Expenses and Recurring Profits, 1960-95

    (Share of sales; four-quarter moving average)