Annual Report on the Japanese

Economy and Public Finance


- No Gains Without Reforms IV -

July 2004

Cabinet Office

Government of Japan

 [Toc]  [Prev]  [Next]  [Annual List

Section 3 Improvement in the Labor Market and Household Sector

The corporate sector is recovering steadily as a result of the economic recovery, with the spread of increasing profits in different industries. Whether this recovery in the corporate sector is spreading to the household sector in the form of increased employment and increased wages is an important point in determining whether the current economic recovery is sustainable. Employment and wages, as of now, have been gradually improving, but still remain weak. This trend in which the recovery in the corporate sector does not spread to the recovery of employment and wages is not limited to only the current recovery phase, but has been apparent since the latter half of the 1990s and possibly reflects structural factors. This section looks at factors behind the lack of growth in employment, reasons the unemployment rate is decreasing despite the stagnant growth in employment, and whether the increase in corporate profits will spread to wages in the future. The section will then discuss factors behind the relatively steady increase in personal consumption with the limited growth of income.
1. Improvement in the Labor Market Despite Remaining Difficulties

Employment by industry, gender and type of employment
    The number of employees, after it weakened in the first half 2003, started to pick up in 2004. However, it has only increased slightly compared with the level it was at in the beginning of 2002 when the current economic recovery began. According to an examination of the employment situation by industry in terms of cumulative increases and decreases in the number of employees from the beginning of 2002 when the economy was in a trough, employment increased only in service industries, which includes medicine and welfare, and education (Figure 1-3-1 (1)). On the other hand, the largest decrease in employment was in manufacturing industries. The number of employees in construction industries and in some tertiary industries such as wholesale, retail, pubs and restaurants also decreased. Regarding the movements of the number of employees by gender, while female employees increased consistently, male employees continued to decrease until around October 2003 and later increased slightly (Figure 1-3-1 (2)). Also, regarding the situation by type of employment, the number of part-time workers increased, and the number of full-time workers decreased by about the same amount as the number of part-time workers increased(3) (Figure 1-3-1 (3)). In sum, it is apparent that with the significant increase in employment in service industries such as health and welfare, female employees increased reflecting the fact that these industries engage a lot of female workers. In addition, the number of part-time workers is increasing especially in tertiary industries.
    The modest growth in employment compared with the past may reflect, in terms of the supply of labor, a reduction in the labor force population as a result of the declining birthrate and aging of the population. On the other hand, given that the supply of labor could be increased considering the large number of people remaining unemployed and the rise in the labor participation rate of females, etc., there is a possibility that modest growth of employment reflects weak employment demand at companies mainly in the manufacturing industry, etc. The following section looks at increases and decreases in such employment opportunities.
Figure 1-3-1 Cumulative Change in Employment Since the Beginning of the Recovery

Business cycles and employment
    If the economic recovery continues hereafter, will employment increase even in industries in which employment does not appear to be increasing now? In order to look at this point, increases and decreases in employment were examined by industry, with the observation period divided into expansion and recession periods (Table 1-3-2)(4). According to the results, there are several findings: (i) employment is increasing in service industries regardless of business cycles (structural gain industries), (ii) employment is decreasing in manufacturing industries, finance and insurance industries and construction industries, etc. regardless of business cycles (structural loss industries), (iii) the effects of business cycles can not be seen in most industries due to the tendency of Japanese companies to maintain employment even when the economy worsens (labor hoarding), and (iv) employment in wholesale and retail industries, etc. is irregular and irrespective against business cycles. Looking at employment share by industry as of May 2004, employment in industries with structural losses made up 40% of all employment, while employment in industries in which there are structural gains made up only approximately 30% of all employment.
Table 1-3-2 Change in Employment by Industry and the Business Cycle
    Looking at the relationship between employment by industry and business cycles over time, the share of structural gain industries decreased gradually from its peak of 80% in the early 1990s to around 60% in the latter half of the 1990s and 30% in 2004 (Appended Figure 1-10). Meanwhile, while the share of employment in structural loss industries was close to zero until the middle of the 1990s, it was over around 40% in the latter half of the 1990s.
Structural changes in job creation/loss
    Here, sequential changes in the job creation capacity in the industries for the latter half of the 1980s to 2001 were examined, by decomposing them into three factors: job creation resulting from business start-ups, changes in employment at existing workplaces, and job losses resulting from business closures (Figure 1-3-3). According to the results, one distinctive feature was that compared to other industries, job creation resulting from business start-ups at all points in time was large in service, wholesale, retail, pubs and restaurants and the transport and communications industries, in which there was relatively steady growth in employment. However, the rates of growth in employment in wholesale, retail, pubs and restaurant, and transport and communication industries decrease on the whole resulting from both a decreasing rate of growth in employment at existing workplaces in the late 1990s and increasing job losses from business closure. Meanwhile, the characteristics of industries in which employment began to fall from the latter half of the 1990s are somewhat different for each industry. In manufacturing industries, job losses resulting from business closures were almost the same at each point in time, but job creation resulting from business start-ups and increases in employment at existing workplaces receded, and this resulted in an overall decrease in employment. In construction industries, and finance and insurance industries, job losses resulting from business closures increased and employment at existing workplaces decreased in the latter half of the 1990s, thus resulting in an overall decrease in employment.
Figure 1-3-3 Job Creation/Destruction Resulting from Business Startups/Closures and Employment Change at Existing Workplace
    In summary, the weak growth in employment in the current recovery period reflects that with the progression of structural losses in employment in the manufacturing, construction, and finance and insurance industries, etc., employment is growing in only a portion of tertiary industries and the job creation capacity in certain tertiary industries is not as sizeable as in the past. Industries in which there are structural losses in employment such as the manufacturing, construction, and finance and insurance industries are faced with the problem of rationalization accompanying shifts of manufacturing hubs overseas, reduction in public works and disposal of non-performing loans, thus making it unlikely that a sustained increase in employment will be retained in these industries even if the economic recovery continues in the future. It thus appears that the degree to which employment can be increased in the future depends on the extent of business start-up increases and job opportunity creation particularly in such industries as the service industry.
Background of declining unemployment rate
    In the previous recovery phase in the 1990s, the unemployment rate did not decrease even though the economy recovered, and as a result, the unemployment rate has consistently continued to increase. In the current recovery phase, however, the unemployment rate began to decrease in the beginning of 2003 and fell to 4.6% in May 2004 from its peak of 5.5%. As considered earlier, the difference between the current recovery phase and the past recovery phase is not significant in terms of the recovery of employment being weak. In order to look at factors behind this current drop in unemployment rate, movements in the flow data for employment, unemployment and non-employment (stopping looking for work and exiting the labor market) were focused on and looked at in addition to the stock data for unemployment(5).
    Specifically, the number of people who moved from employment or non-employment (a state of having exited the labor market) to unemployment and the number of people who moved from unemployment to either employment or non-employment were first calculated for each point in time. These flow figures were then divided by stock figures for employment, unemployment and non-employment, and transition probabilities, which show the probability of a person moving from one state to the other, were calculated (Figure 1-3-4). According to the results of these calculations, there has been no significant change from the 1990s to recently in the actual probability that a person will move from unemployment to employment or non-employment. Specifically, the probability that a person will move to employment from unemployment decreased gradually during the mid- to late 1990s, rose slightly in 2002, and was flat after that(6). The probability that a person will move from unemployment to non-employment was for the most part flat. Meanwhile, the probability that a person will enter a state of unemployment from a state of employment or non-employment rose rapidly in the latter half of the 1990s and has been decreasing clearly since around 2002 and 2003. Factors behind the decline in the unemployment rate since the beginning of 2003 despite the low growth in employment include (i) a decline in the probability of a person newly entering a state of unemployment from a state of employment as a result of the large-scale reduction in personnel through restructuring by companies having moved past its peak, and (ii) a decline in the probability of a person moving from non-employment to unemployment though such a probability had been raised by housewives and other people who had not been working to look for work. This point can be clearly indicated by the trends by gender. A distinctive feature among men is that the probability of a person moving from employment to unemployment has been declining steadily since the middle of 2002, and among women, the probability of a person moving from non-employment to unemployment has been falling since the beginning of 2003 (Appended Figure 1-11).
Figure 1-3-4 Labour-force Flow (Transition Probability)

Column 1-1

Why did the unemployment rate not rise as much as expected despite the disposal of non-performing loans?
    In its Emergency Economic Package in April 2001, the Japanese government announced a policy in which non-performing loans by major banks to firms that have already gone bankrupt or are in danger of going bankrupt will in principle be directly disposed of (off-balanced) within three years of their occurrence. At the time, there were concerns that this disposal of non-performing loans could result in a significant increase in the number of corporate bankruptcies and an increase in the number of people unemployed. For example, in June 2001, the Cabinet Office announced estimated results showing that between 130,000 and 190,000 people could become unemployed as a result of the final disposal of 12.7 trillion yen (at the end of September 30, 2000) worth of non-performing loans by major banks to firms that had already gone bankrupt or were in danger of going bankrupt. The following are the results of an ex-post verification looking at the extent to which unemployment increased due to the disposal of non-performing loans (see Kato, Fujiwara and Fujimoto (2003) for details).
    The final disposal of non-performing loans by major banks to firms that had already gone bankrupt or were in danger of going bankrupt in FY2002 in fact amounted to 11.7 trillion yen. This was not significantly different from the figure announced in 2001 of 12.7 trillion yen in non-performing loans to firms that had already gone bankrupt or were in danger of going bankrupt. Based on statistics for bankrupt companies, the number of employees at bankrupt companies which borrowed from major banks was calculated to be 64,000 for companies which underwent liquidation-type bankruptcies and 36,000 for companies which underwent rehabilitation-type bankruptcies. The number of employees at companies rehabilitated in a privately-arranged scheme which underwent debt forgiveness or debt-for-equity swaps was between 100,000 and 360,000 depending on the method of calculation. Assuming that the turnover rate at companies which underwent a liquidation-type bankruptcy is 100%, 38.1% at companies which underwent a rehabilitation-type bankruptcy and 20% at companies which underwent a private arrangement, the number of displaced workers was between 100,000 and 150,000. Assuming that about under 50% of these displaced workers became unemployed, the number of people who lost their jobs due to non-performing loan disposal in FY2002 can be calculated as having been between about 50,000 and 70,000(7).
    Reasons that the unemployment due to non-performing loan disposal did not increase as much as estimated at the initial stage of the non-performing loan disposal include the fact that with the enhancement of schemes to rehabilitate companies, the actual number of bankruptcies was significantly lower than assumed and the rehabilitation-type of non-performing loan disposal increased more than the liquidation-type (Column Figure 1-1).
Column Figure 1-1 Increasing Number of Reorganization-Type Disposals Due To Bankruptcies

Youth unemployment remaining at a high level
    While the unemployment rate has been decreasing somewhat, the number of long-term unemployed, or people who have been out of work for more than one year, continues to be at a high level, reaching 1.12 million in the first quarter of 2004. The number of long-term unemployed as a proportion of all unemployed, despite a slight drop, continues to be approximately one-third (Figure 1-3-5). This hovering of long-term unemployment at a high level partially reflects the fact that the number of jobs is not increasing as much as the number of job offers. There is an undeniable possibility that it is due to increases in mismatches between jobseekers and job offers. With regard to jobseekers, however, it may be difficult to simply compare the current makeup of job openings with those in the past given the significant increase recently in job offers of contract and dispatch jobs, etc. In addition to these changes, it is difficult to measures the degree of mismatch using only macro-statistics, thus making it difficult to obtain an overall judgment regarding the degree of unemployment which is due to mismatch(8). The status of corporate level vacancies in the survey on employment trends was examined for reference. Looking at vacancy rate (the number of job openings for filling vacancies in a particular job position divided by the number of workers in a job position) by job position, the vacancy rate declined for almost all job positions from the latter half of the 1990s to 2003, while the vacancy rate of particular job positions during that time did not appear to rise due to mismatch (Appended Figure 1-12).
Figure 1-3-5 Number of Those Remaining Unemployed for One Year or Longer
    Returning to the issue of long-term unemployment, by age, the largest portion of unemployed are between the ages of 25 and 34, and even though unemployment is falling slightly on the whole, the number of unemployed in this age group still continues to rise (Figure 1-3-6)(9). There are two factors behind this increase in the number of unemployed youth: labor demand on the side of companies and the attitude toward job-seeking by the unemployed youth.
Figure 1-3-6 Number of Those Remaining Unemployed for One Year or Longer (By Age)
    First, regarding factors on the labor demand side, opportunities for companies to employ the youth are decreasing. This reflects the fact that companies are holding back on recruitment when planning adjustments to personnel and reducing the hiring of new graduates as they seek more people who can be utilized right away without training, etc. Regarding the composition of employees at companies by age, the proportion of middle-aged and elderly workers is increasing and the proportion of young workers is decreasing. This trend is particularly apparent at large companies, which adopt life-long employment (Appended Figure 1-13). An empirical analysis by Genda (2000) shows that the higher the proportion of employees aged 45 and older at a company, the more the company constrains its hiring of new graduates.
    Second, regarding factors on the side of the unemployed youth, a major aspect is mismatch whereby there is a lack of work which meets the conditions of young people. Another factor appears to be that the unemployment rate of unmarried youth living with their parents is higher than that of young people overall, and this group is less effective in job search (Appended Figure 1-14).
    Regarding the forms of employment among the youth, the number of young people who work at non-regular jobs as freeters (people preferring not to take a regular job) has been increasing, but the probability of young people moving from non-regular jobs to regular jobs is low, and there is a possibility that more young people could become permanently categorized as freeters (White Paper on the National Lifestyle Fiscal Year 2003).
    In any event, with young people being unemployed or in a non-regular job for prolonged periods of time given their limited work experience, the formation of human capital including the accumulation of professional know-how will be insufficient. As a result, it will be difficult for the youth to avoid further unemployment and non-regular work, and it will be necessary to strengthen employment measures for the youth including support for the formation of human capital through improving job consciousness, strengthening job matching functions and job training, etc.
Effect of improved company profits on wages
    Wages continue to be weak despite improvements in corporate profits. Factors behind why wages are not increasing relative to the economic recovery include that (i) companies are still continuing to be cautious about increasing their labor costs even though restructuring has pretty much come to a halt, and (ii) there has been a rise in the proportion of part-time workers whose salaries are relatively low compared with those of full-time workers.
    First, regarding the stance of companies on raising wages, in terms of employment, companies have started to halt restructuring taking the form of reduction in the number of employees. In terms of wages, the proportion of companies which have not made revisions to wage rates such as pay-scale hikes has been increasing. However, while many companies have been restraining basic salaries, they distribute the gain from profit through increasing bonuses (Appended Figure 1-15). This attempt by companies to restrain wage increases and suppress labor costs can also be viewed in terms of movements in labor share in the economy as a whole. Labor share is determined by the relative size of productivity and real wages. With productivity dropping in 2002, both nominal and real wages fell significantly, resulting in a decline in labor share by about 1% relative to GDP (Appended Figure 1-16). Labor share fell further in 2003, reflecting the fact that the growth in real wages was restrained to 1%, staying much below the labor productivity growth of 2.4%.
    Second, factors behind movements in contractual cash earnings, which include overtime pay and base wages, were broken down in order to look at the effect of rises in the proportion of part-time workers. Although the wages of full-time workers began to rise at the end of 2002, a rise of the proportion of part-time workers whose wages are relatively low pushed down the growth in contractual cash earnings overall (Figure 1-3-7). The phenomenon of wages being restrained by rises in the proportion of part-time workers is more prominent in the current recovery period than it was during the previous recovery period.
Figure 1-3-7 Breakdown of Contractual Cash Earnings

Will wages rise in the future?
    Will growth in wages continue to be low even if economic recovery continues in the future? Companies basically appear to continue to be cautious about raising wages. In the 2004 spring labor offensive, increases in wages including contractual cash earnings granted to employees were kept at almost the same level as in the previous year, and bonuses were raised, reflecting a recovery in the financial results of companies. As such, growth in base wages is anticipated to continue to be restrained, and if the economic recovery continues, employee wages are expected to increase gradually in the form of overtime pay and bonuses. Comparing the relationship between productivity increases by industry and growth in real wages in 2002 and 2003, there was a significant increase in 2003 in the number of industries in which real wages increased with rising productivity (Figure 1-3-8). Thus, if the economic recovery further expands, it appears that the number of industries in which there are increases in both productivity and wages will increase further.
Figure 1-3-8 Relationship between Labor Productivity and Real Wage Growth by Industry
    Meanwhile, with regard to the rise in the proportion of part-time workers employed, companies will continue to take a positive approach toward the employment of part-time workers, and there is thus a possibility that increases in compensation for employees will continue to be restrained as a result of a progressive rise in the proportion of part-time workers whose wages are relatively low. According to the Cabinet Office Annual Survey of Corporate Behavior carried out in January 2004, companies which responded to the survey had increased the number of their part-time and temporary workers by an average of 2.2% over the past three years and indicated that they plan to increase their part-time and temporary workers at about the same pace over the subsequent three years. Companies reduced the number of their full-time workers by 3.4% over the past three years. Although the pace of the reduction of full-time workers will slow, with companies planning to reduce their full-time workers by 0.8%, the employment of full-time workers appears to continue to be suppressed.
    In summary, although the number of companies at which wages are gradually increasing is expected to increase as a result of the continued economic recovery, growth in compensation of employees in the economy as a whole is expected to be modest reflecting a progressive rise in the proportion of part-time workers.
2. Stable Demand in Household Sector

Modest improvement in private consumption
    Over the past few years, household income has been weak while personal consumption has been relatively strong. This reflects the fact that consumption was stabilized by a fall in the savings rate in accordance with a decrease in household income. The change in the savings rate appears to reflect long-term trends such as an increase in the proportion of elderly households which are using their savings from the perspective of life cycles, and an accumulation of financial asset balance, etc. (Appended Figure 1-17). Meanwhile, consumption behavior has not changed significantly with periodic increases or decreases in income. As a result, it appears that savings has changed so as to act to stabilize consumption.
    In order to distinguish these long-term trends in consumption and savings rates from consumption smoothing (consumption equalization), which responds to short-term changes in household income, it was assumed that there is a long-term equilibrium among consumption, disposable income, the proportion of elderly people in the population and financial assets. The degree to which actual consumption diverges from this long-run equilibrium was examined below (Figure 1-3-9). According to the results of this estimation, actual consumption was slightly higher than the long-term equilibrium level around 1996 and 1997. Following this, consumption was lower than the long-term equilibrium level from 1998 to about 2000, partially as a result of a financial crisis and employment uncertainty accompanied by a rise in the unemployment rate, etc. In 2001, consumption was higher than the long-term equilibrium level, and there was a significant decline in the savings rate. This appears to reflect not only a fluctuation in statistics accompanying a large number of postal savings accounts reaching maturity, but also reactions to declines in the previous period and the smoothing of consumption during a periodic decline in household income(10). Consumption in 2002 was comparably not as high above the long-term equilibrium level, and the level of consumption is gradually converging toward a long-term equilibrium determined by disposable income, the proportion of elderly people and financial assets. In extending the estimate by making certain assumptions and calculating disposable income, it appears that consumption in 2003 is falling again toward the long-term equilibrium level. Recent movements in consumption from year to year have thus been generally consistent with long-term factors such as changes in household income and makeup of the population. The strengthening of consumption from the end of 2003 to the beginning of 2004, however, cannot be sufficiently accounted for by long-term factors alone, and factors mentioned below such as the improved employment situation also appear to have an effect.
Figure 1-3-9 Trends in Consumption, Savings Rates and Income

Recovery of consumer confidence through improved employment situation and new products contributing to creating new demand for consumption
    Consumption on the whole appears to be recovering since the latter half of 2003, and this seems to reflect (i) a recovery in consumer confidence due to the improved employment situation including decreases in the unemployment rates, and (ii) latent demand being uncovered by the emergence of new products such as digital electronics.
    First, regarding consumer confidence, restructuring at companies, as mentioned earlier, is coming to a halt, and the number of involuntarily unemployed is decreasing. As a result, the outlook of consumers regarding the future direction of employment has improved, and this is contributing significantly to improving consumer confidence (Figure 1-3-10). There is thus a possibility that consumption which had previously been held off is emerging as a result of a sense of relief regarding the future direction of employment.
Figure 1-3-10 Number of Involuntary Unemployed Decreased and Confidence Improved
    Second, dividing consumption expenditure into selective expenditure which includes eating out, travel, DVDs and personal computers, and basic expenditure which includes food, and lighting and heating expenses, the contribution of selective expenditure appears to account for most of the increase in expenditure (Figure 1-3-11). In addition, sales of so-called digital electronics such as flat-screen televisions, DVD players and digital cameras, have been strong. A consumer survey was carried out to investigate the types of people who purchase digital electronics and whether the characteristics of digital electronics differ from those of other goods (see Appended Figure 1-3 for further details). The questionnaire involved 2,000 people throughout Japan, and valid responses were received from approximately 1,200 people. The main findings of the questionnaire are as follows.
Figure 1-3-11 Breakdown of Consumption Expenditure

(1) Types of purchasers
    More than 50% of the respondents said that they purchased or planned to purchase a digital camera, about 40% said that they purchased or planned to purchase a DVD player, and about 10% said that they purchased or planned to purchase a flat-screen television (Appended Figure 1-3). One characteristic of these purchasers of digital electronics is that they were in a relatively wide range of age groups. The overall popularity of flat-screen televisions is still low, but flat-screen televisions were more popular among elderly people, who have relatively more free time, than among young people (Figure 1-3-12).
Figure 1-3-12 Digital Electronics Purchase by Age

(2) Demand creation effect of digital electronics
    In order to investigate the extent to which digital electronics created new consumer demand, the respondents were asked about their habits in replacing old models. The percentage of respondents who answered "I could have still used the old model, but I bought a new one earlier than usual" and "I bought an entirely new one that I did not own before" was 80% of the total purchasers for digital cameras, over 70% for DVD players and 60% for flat-screen televisions (Figure 1-3-13). Considering that this figure is around 30% for purchasers of automobiles, white goods and other durable goods, these results seem to show that new products had a considerable effect in creating demand. In addition, when respondents were asked if purchasing digital electronics would reduce their spending in other areas, around 50% to 60% of respondents answered that they did not think that it would make them save in other areas, suggesting that the correlation of substitution and consumption of other goods is not so large.
Figure 1-3-13 Net Effect on Demand Created by Digital Electronics

Effect of digital electronics on production
    Strong sales of digital electronics have also led to positive movements in production. Comparing current trends in the production of electric machinery, information and communications machinery and electronic parts and devices with those in 2000, when the last peak in the economy occurred, production of information and communications machinery is not at the level as during the previous period and that of electric machinery has grown slightly. Electronic parts and devices and household electronic machinery, a classification of information and communications machinery, which are related to digital electronics, however, are significantly higher than in the last period (Figure 1-3-14). One characteristic of digital electronics in terms of production is that production ripple effects in the field of electronic devices are strong due to the high concentration of parts and Japanese companies have a high share in parts for digital electronics. As a result, this virtuous circle for Japanese companies is expected to continue for the time being. In addition, it appears that a high percentage of products such as digital electronics, which are highly technologically-intensive and have a high level of value added, are produced domestically.
Figure 1-3-14 Trends in Digital Electronics
    According to the FY2003 White Paper on Manufacturing Infrastructure, the production ripple effect (including related industries) of digital electronics in 2003 was approximately 7 trillion yen, and it appears that the added-value trigger effect on related industries was higher than in other industries.
Modest increase in housing investment
    Housing construction starts have been decreasing since the latter half of the 1990s, but were relatively firm in FY2003, growing 2.5% from the previous year (Figure 1-3-15 (1)). According to a calculation of the contributions to this by region and usage, most of the growth in housing construction nationally can be attributed to an increase in houses for sale and houses for rent in Southern Kanto area, including Tokyo (Figure 1-3-15 (2)). This increase in housing construction in the Tokyo area reflects increased construction of apartments and houses for rent due to an increase in residences converted from corporate sites in relatively inner city areas, along with an inflow of the population into the downtown area. Land prices have clearly stopped falling in the Tokyo wards and the surrounding areas, but land prices in local areas continue to decline (Figure 1-3-16). Meanwhile, with regard to the ability of individuals to acquire housing in terms of the income and interest environments, given that improvements in the corporate sector have not been reflected in income, the ability to acquire housing has remained almost flat (Appended Figure 1-18). Regarding future directions, if the income environment gradually improves with sustained economic recovery, housing construction starts can be expected to continue to move steadily. Regarding the situation by region, housing starts will likely continue to increase in the Tokyo area for the time being, as reflected in land prices, etc.
Figure 1-3-15 New Housing Starts
Figure 1-3-16 Official Land Price (Residential)

 [Toc]  [Prev]  [Next]  [Annual List