UNIVERSE OF 
 * BOOKSTACKS 
 
Digitized by the Internet Archive 
 
 in 2011 with funding from 
 
 University of Illinois Urbana-Champaign 
 
 http://www.archive.org/details/economicgrowthla909tair 
 
FACULTY WORKING 
 PAPER NO. 909 
 
 Economic Growth, Labor Productivity and 
 Employment Adjustment in Japan 
 
 Koji Taira 
 
 DEC 
 
 The 
 
 »9f 
 
 oonege of Commence and Business Administration 
 Bureau of Economic and Business Research 
 University of Illinois. Urbana-Champaign 
 
FACULTY WORKING PAPER No. 909 
 College of Commerce and Business Administration 
 University of Illinois at Urbana-.Charapaign 
 October 1982 
 
 Economic Growth, Labor Productivity and 
 Employment Adjustment in Japan 
 
 Koji Taira, Professor 
 Department of Economics 
 
 Working Paper: Not to be quoted 
 without the author's permission 
 
Abstract 
 Japan's preponderant emphasis on the fastest possible improvement 
 of labor productivity in manufacturing, unless the world keeps absorbing 
 Japanese exports at an increasing rate, may well reduce Japan's own 
 overall economic growth by shifting labor from high-productivity 
 manufacturing to low-productivity sectors. The first part of this paper 
 presents statistics to indicate where Japan stands relative to the U.S. 
 by real income and labor productivity. The second part examines employ- 
 ment adjustment in Japanese manufacturing and suggests that Japan's 
 success in the reduction of manufacturing employment has been due to the 
 weakness of job security, almost ruthless managerial rationality combined 
 with superb techniques of conflict minimization, understanding responses 
 of labor unions, and workers' own work-biased work-leisure preference 
 functions. 
 
Economic growth and productivity growth 
 
 Table assembles some comparative statistics on the growth of 
 real income and labor productivity useful for assessing the relative 
 standing of the U.S. and Japan. 
 
 (Table about here) 
 
 What is interesting as revealed by the columns of GNP growth rates 
 is that after the flare-up of the Japanese GNP growth rate in 1972, which 
 was a relatively peaceful year between the Nixon Shock of 1971 and the 
 OPEC Shock of 1973, Japanese economic growth has slowed to low rates, 
 which for two consecutive years of 1976 and 1977 even fell below the U.S. 
 growth rates. It is generally believed that with the OPEC Shock Japan 
 entered a new era of slow growth. Statistical data seem to bear that 
 out. The second OPEC Shock stretched out from early 1979 to the end 
 of 1980, doubling the crude oil prices over the period. Japan managed 
 this crisis much better than the earlier one and certainly much better 
 than the U.S. did. The important lesson, however, is that other 
 things (especially economic policy) being the same, there is no reason 
 why the U.S. cannot attain the rate of economic growth equal to 
 Japan's in the world of the 1980s. Potential growth rates are pro- 
 bably similar in Japan and the U.S. now. But the greater likelihood 
 of policy mistakes in the U.S. may translate the same potential into 
 poorer performance. 
 
 The column of Japan's per capita output relative to that of the 
 U.S. shows that there still Is a considerable difference between Japan 
 and the U.S. The relativity may be subject to some inherent bias due 
 
-2- 
 
 to the particular method of estimation and comparison used by the 
 authors. At present, we do not know whether the bias exists at all or 
 whether, if it does, it works against the U.S. or Japan. Japan's per 
 capita output at about 60 percent of that of the U.S. in the early 
 1970s seems intuitively acceptable. The figure shown for 1980 in 
 parentheses is obtained by applying to the 1970 figure the relative 
 index of real GNP per capita reported for Japan and the U.S. (index of 
 real GNP per capita for Japan divided by the index of real GNP per 
 capita for the U.S.). For 1980, Japan's per capita output comes to be 
 more than 30 percent below that of the U.S. This difference seems 
 reasonable in the light of other numbers shown in Table 0. 
 
 The rest of Table refers to manufacturing, except for the last 
 column labelled "all industries." All of these relative estimates are 
 full of technical and conceptual problems. They should be taken only 
 as illustrative, although better estimates are not available. 
 Surprisingly, however, they seem to show a degree of consistency among 
 them, indicating that they must be measuring the same thing — labor 
 productivity in manufacturing. A systematic bias in favor of Japan is 
 endemic to the method used here — matching of comparable industries 
 which leaves out industries which exist in one country but do not in 
 another or industries whose products are not measured in common units. 
 Also, the weights of the "matched industries" are different between 
 countries. The original estimate by Kenzo Yukizawa for 1967 on which 
 the Labor Department adjustments were grafted is appraised to cover 25 
 percent of U.S. gross value added and 32 percent of Japanese net value 
 added in manufacturing, as well as 21 percent of U.S. employment and 
 
-3- 
 
 24 percent of Japanese employment In manufacturing. This gives rise 
 to the question of representativeness of the industries covered for 
 international comparison and to the impression that the estimate 
 
 covers a higher proportion of relatively efficient industries in Japan 
 
 2 
 than in the United States. But there is no way of adjusting for this 
 
 bias. One can merely say that Japanese productivity may not be really 
 as high as the figures suggest. The figures in parentheses for 1979 
 were extensions of earlier indices by relative indices of productivity 
 from another source as indicated in the notes to the table. It is 
 significant that by U.S. price weights, Japanese productivity had 
 caught up with American. 
 
 The estimate of Japan's physical labor productivity relative to 
 that of the U.S. for 1977 is done by Japan's Labor Ministry. This 
 covers products which are common to both countries and measured and 
 reported in physical units. The use of physical units seems to 
 restrict the scope of "matched industries." The number of production 
 workers in these industries, according to our own check on the basis 
 of censuses of manufactures is 15.1 percent of all production workers 
 in manufacturing for the U.S. and 16.2 percent for Japan. The value- 
 added productivity of the "matched industries" is roughly on a par 
 with that of all manufacturing in the U.S., but 22 percent higher than 
 that of all manufacturing in Japan. It appears therefore that the 
 Labor Ministry estimates results in a favorable showing for Japan than 
 would have been the case if all manufacturing had been covered 
 (although how this can be done is technically unsolvable as in the 
 case of the U.S. Labor Department work). Furthermore, it is generally 
 
-4- 
 
 believed that value added is a smaller proportion of the value of 
 shipments in Japan than in the U.S. Thus, the comaprison of physical 
 productivity would tend to favor Japan more than the U. S, It is 
 instructive, however, that both the Labor Ministry figures and the 
 U.S. Labor Department figures for 1977 and 1978 are close to each 
 other. This warrants a conclusion that in the late 1970s Japanese 
 labor productivity in manufacturing had attained rough parity with 
 that of the U.S. and that since Japan was consistently gaining on the 
 U.S. all the while, Japanese labor productivity in manufacturing may 
 well have surged past that of the U.S. at some time in the early 1980s. 
 
 The last two columns are based on a press report on a study done 
 at the Japanese Productivity Center. Until we see a full report on 
 the study, we cannot say much about the nature of these estimates. 
 The methodology is apparently the same as the previous estimates by 
 Yukizawa and the Labor Ministry. The proximity of the figure for 
 manufacturing to the other estimates strengthens the general conclu- 
 sion reached in the preceding paragraph. But, what is more important 
 is the fact that in Japan, labor productivity in manufacturing is far 
 higher than that in all industries. Thus, even if it is true that 
 Japan is as efficient sa the U.S. in manufacturing one cannot say that 
 the economy as a whole is similarly efficient. The relative labor 
 productivity in all industries is strikingly similar to the relative 
 per capita output for 1980 in the third column. This then gives rise 
 to a tentative conclusion regarding Japan's economic level vis-a-vis 
 the U.S. Although manufacturing efficiency is on a par with the U.S. 
 
-5- 
 
 or even rising above the U.S. , the general economic well-being of 
 Japan still lags behind that of the U.S. in the order of 30 to 40 per- 
 cent. 
 
 What makes manufacturing productivity surge ahead in that manner 
 leaving other industries far behind? This is an interesting question, 
 but a complex one on which we have no time to dwell. Compare this 
 pattern of uneven development in productivity with the United States' 
 inter-industrially rather balanced growth of productivity, and one 
 sense another potentially interesting question, which is: under which 
 pattern of inter-sectoral productivity growth, balanced or unbalanced, 
 is real per capita GNP likely to grow faster in the long run? Japan's 
 economic growth during the period of fast growth was supported by 
 growth-inducing resource allocation that the differential structure 
 had produced. This was the classic mechanism of shifting manpower 
 from low-growth, low-productivity industries to high-growth, high- 
 productivity industries. But that is precisely the problem for Japan 
 under the economic structure that has evolved since the early 1970s. 
 Manufacturing, the most dynamic high-productivity sector, has ceased 
 to absorb more manpower or, worse, has been shedding manpower con- 
 sistently over several years. The labor force that manufacturing eli- 
 minates or cannot absorb must be absorbed in lower-productivity sec- 
 tors. To generalize from this, the differential structure now alloca- 
 tes manpower from high-productivity to low-productivity sectors, while 
 the growth rate of total output in the manufacturing sector has slowed 
 down compared with earlier years, making the share of manufacturing in 
 GNP smaller. Allocative processes like these are bound to slow down 
 
-6- 
 
 the growth of overall productivity of the entire economy, making it lag 
 further behind manufacturing productivity. 
 
 The ultimate of manpower shedding in manufacturing is perfect auto- 
 mation or robotization which the Japanese actually proclaim as their 
 goal. This means that 25 percent or so of the Japanese labor force 
 which is still engaged in manufacturing has to be shifted to other low- 
 productivity sectors. If economic logic works as it should in the real 
 world, the greater availability of labor in these sectors should 
 encourage labor intensive processes and retard productivity growth. In 
 the meantime, the growth of manufacturing output is limited by the level 
 of income and the income elasticity of demand for industrial products 
 at home and abroad. Since personal incomes are earned only from the 
 low-productivity, low-growth sectors, the domestic demand for manufac- 
 turing output can grow only slowly. This results in an increasing 
 underutilization of robots; i.e., an increasing excess capacity in 
 manufacturing. The rate of economic growth may still be positive, but 
 whether it is even as high as the average growth rate of the 1970s is a 
 good question. The growth rate can be kept high if the foreign demand 
 for Japanese manufactures grows faster than the growth of labor produc- 
 tivity in Japanese manufacturing. An innovative foreign marketing of 
 the Japanese and a boundless good will of foreign consumers may for a 
 
 while sustain Japanese growth as in recent years. That it is too good 
 
 3 
 to last forever is the biggest worry of the Japanese. 
 
 This scenario is of course extreme, but it brings out one impor- 
 tant question. Are the Japanese really serious about shedding manu- 
 facturing labor? The Japanese experience of employment adjustment in 
 manufacturing may suggest at least a partial answer to this question. 
 
-7- 
 
 Footnotes 
 
 For useful commentary on this and other methods of international 
 comparison of economic growth, see Hugh Patrick, "The Future of the 
 Japanese Economy: Output and Labor Productivity," Journal of Japanese 
 Studies , 3,2 (Summer 1977): 219-249. Unfortunately, almost exclusive 
 attention on the comparison of growth tends to neglect the comparison 
 of levels of income or productivity at a given point in time. 
 
 2 
 U.S. Department of Labor, Comparative Growth in Manufacturing 
 
 Productivity and Labor Costs in Selected Industrialized Countries , pp. 
 
 18-19. 
 
 3 
 Japan's "export-led" or "export-biased" growth is the technical 
 
 heading under which this sort of discussion is carried out by econo- 
 mists. In recent years, many economists have reduced the role of 
 exports as a leader of Japanese growth, moderating as a consequence 
 their view of Japan's potential growth rates. An illuminating contri- 
 bution to the discussion is Shinkai Yoichi, "Patterns of American and 
 Japanese Growth and Productivity: A Japanese Perspective," Japan 
 Quarterly , 27, 3 (July-September 1980): 358-375. 
 
 D/122 
 
a 
 
 3 
 T3 
 O 
 U 
 
 a. 
 
 
 >> 
 
 
 4-1 r-~ 
 
 
 T-l • 
 
 
 > CO 
 
 
 ■H • 
 
 a 
 
 U => 
 
 3 
 
 o 
 
 H 
 
 3 O 
 
 cO 
 
 T3 4J 
 
 > 
 
 o 
 
 
 u OJ 
 
 00 
 
 a. > 
 
 M 
 
 ■H 
 
 3 
 
 T3 4-1 
 
 cfl 
 
 a) to 
 
 a-o h 
 
 co 
 
 •a m 
 
 -: 
 
 cfl u 
 
 < 3 
 
 -a 
 
 c 
 
 
 60J 
 
 
 c 
 
 1 
 
 ■H 
 
 - 
 
 U 
 
 c 
 
 3 
 
 a 
 
 j-i 
 
 s 
 
 CJ 
 
 
 CO 
 
 
 14H 
 
 co 
 
 en 
 con 
 
 s-i 
 o 
 
 T3 
 
 c 
 CO 
 
 4J 
 
 3 
 
 a. 
 
 4J 
 
 3 
 O 
 
 CO 
 
 4-1 
 
 •H 
 
 a. 
 co 
 o 
 
 
 ►> 
 
 
 j_ivO 
 
 rH 
 
 •H • 
 
 CO 
 
 > CO 
 
 a 
 
 •H • 
 
 ■H 
 
 4-1 S 
 
 01 
 
 o 
 
 >, 
 
 3 O 
 
 42 
 
 •73 4_i 
 
 a 
 
 O 
 
 
 Ij CU 
 
 01 
 
 a > 
 
 — 
 
 •H 
 
 c 
 
 S-I 4-1 
 
 eo 
 
 O CO 
 
 a. ja ih 
 
 CO 
 
 CO <U 
 
 ^ 
 
 iH U 
 
 u 
 
 U 3 
 
 <u o 
 
 O. JS 
 
 (H 
 
 CD 
 
 S-i M 
 
 01 O 
 
 O. 3 
 
 ON 
 00 
 
 on 
 
 o 
 
 w 
 -J 
 oo 
 
 -4 
 
 a 
 
 09 
 
 CD 
 
 CO 
 
 3 
 
 o 
 u 
 
 bO 
 
 tu 
 Z 
 
 o 
 
 3 
 O 
 
 CD 
 O 
 
 
 
 
 11 
 
 
 
 
 CJ 03 
 
 
 
 
 •H 4-1 
 
 
 OJ 
 
 
 u x: 
 
 
 > 
 
 
 S— 60 
 
 
 ■H 
 
 
 •H 
 
 60 
 
 4-1 
 
 
 • 1) 
 
 S 
 
 CO 
 
 
 to 3 
 
 ■H 
 
 H 
 
 
 • 
 
 IJ 
 
 0) 
 
 
 D 
 
 3 
 
 M 
 
 
 
 4-1 
 
 
 
 
 CJ 
 
 h 
 
 
 
 CO 
 
 3 
 
 
 OJ 
 
 14-1 
 
 O 
 
 
 CJ 
 
 3 
 
 X! 
 
 
 •H 
 
 C 
 
 
 
 M 
 
 CO 
 
 h 
 
 
 O. CO 
 
 S 
 
 <U 
 
 
 4-1 
 
 
 Of* 
 
 cu x 
 
 01 
 
 
 • 
 
 03 60 
 
 — 
 
 4-1 
 
 CO 
 
 0) T-t 
 
 c 
 
 3 
 
 • 
 
 C CD 
 
 cfl 
 
 O.S 
 
 CO 3 
 
 a. 
 
 4-1 
 
 
 a. 
 
 CO 
 
 3 
 
 
 
 CO 
 
 ^ 
 
 O 
 
 4-1 
 
 T> 
 
 m cm o co r-. 
 
 r*» o oo <r ao 
 
 \0 *0 O N [^ 
 
 CM 
 X> 
 I — 
 
 o 
 
 m^- 0<t H 
 
 on 
 
 on 
 
 s \o co en rs 
 
 <r 
 
 00 
 
 in in i/> ic vo 
 
 ■X) 
 
 ao 
 
 
 • 
 
 • 
 
 cfl 
 
 4-1 
 Cfl 
 
 CU 
 
 > 
 
 4-1 
 
 Cfl 
 
 U 
 CO 
 
 a. 
 
 io a > • 
 
 CO 4J i-l CO 
 
 C O 3 4-1 • 
 
 co a eg 3 
 
 a ^ uh 
 
 to <u 3 ai o 
 
 !-3 O. O M 4-1 
 
 n Hifl n 
 00 O H cni 
 
 l/l O \D \0 
 
 co 
 
 00 
 
 CO 
 
 ! 
 
 a 
 
 cfl 
 fX 
 
 CO 
 
 >-3 
 
 4-> 
 
 c 
 
 
 
 
 
 
 
 
 CO 
 
 cfl 
 
 CO 
 
 <n 
 
 rv en cn 
 
 X> i-H CO H 
 
 co 
 
 r~~ 
 
 r^ 
 
 Cd 
 
 a, 
 
 
 
 
 
 
 
 
 
 cfl 
 
 00 
 
 in 
 
 on in 
 
 co u~> m m 
 
 uo 
 
 co 
 
 CM 
 
 X 
 
 *-3 
 
 
 
 1 
 
 
 
 
 
 4-1 
 
 
 
 
 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 • 
 
 
 
 
 
 
 
 
 u 
 
 <l 
 
 
 
 
 
 
 
 
 CJ 
 
 • 
 
 CO 
 
 CM 
 
 <r 
 
 P^ CO \D 
 
 w<f into 
 
 CM 
 
 CM 
 
 ON 
 
 3-1 
 
 • 
 
 
 en 
 
 m m 
 
 -4 u-, uo <r 
 
 CO 
 
 
 —J 
 
 2 
 
 p 
 
 1 
 
 
 i 
 
 I 
 
 
 
 
 r O 
 
 
 
 
 
 
 
 1 
 
 
 O HMn-j 
 r^ r-~ r~. r-^ r^ 
 
 ON ON ON ON ON 
 
 in vO N CO ON O rH 
 
 i — r~~ i — r~~ i — oo oo 
 
 ON CON ON ON ON ON ON 
 
-9- 
 
 Notes to Table 
 
 •'■U.S. Government, Economic Report of the President (January 1982); 
 Japanese Government, Economic White Paper . 
 
 ^Laurits R. Christensen, Dianne Cummings , and Dale W. Jorgenson, 
 "Relative Productivity Levels, 1947-1973: An International 
 Comparison" (a working paper) (Social Systems Research Institute, 
 University of Wisconsin, Madison, 1980). 
 
 ■^Obtained by multiplying the Christensen-Cummings-Jorgenson figure 
 for 1970 by the Japan-U.S. ratio of per capita GNP growth rates between 
 1970 and 1980 calculated from data in sources noted in footnote 1. 
 
 4 U.S. Department of Labor, Comparative Growth in Manufacturing 
 Productivity and Labor Costs in Selected Industrialized Countries (BLS 
 #1958, 1977), p. 18. The benchmark study of the U.S. -Japan ratio of 
 productivity is that by Kenzo Yukizawa for 1967. BLS adjusted for 
 other years, including the years shown here, by applying the BLS pro- 
 ductivity growth rates. 
 
 ^Obtained by multiplying the BLS 1970 figures by the Japan/U.S. 
 productivity growth rate ratio between 1970 and 1979 based on Arthur 
 Neef and Patricia Capdevielle, "International Comparisons of 
 Productivity and Labor Costs," Monthly Labor Review , 103, 12 
 (December 1980): 36. " 
 
 6 Ministry of Labor, Labor White Paper (1981). 
 
 Report of the Japan Productivity Center as announced in The Japan 
 Economic Journal (July 20, 1982). 
 
Employment Adjustment in Japanese Manufacturing 
 
 In manufacturing as a whole, the reduction of output during the 
 1973-75 recession was accompanied by a sharp reduction of hours worked 
 per worker and a moderate decline in the number of workers employed. 
 With the increase in output after 1975, hours worked per worker also 
 increased, but the decline in employment continued for some more years 
 until 1979 (see Figure 1). The reversal in the movement of hours 
 worked per worker also seems to go counter to the long run trend of 
 working hours as may be discerned in Figure 1. Manufacturing 
 employment obviously peaked in 1971, suggesting a considerable discon- 
 tinuity in conditions surrounding employment between the 1960s and 
 1970s. 
 
 (Figure 1 about here) 
 
 With the worsening labor shortage during the late 1960s, the 
 rising demand for a better quality of life and a stronger pollution 
 control at the same time, and the re-valuation of the yen in the early 
 1970s, Japanese manufacturing firms were already faced with the need 
 for drastic cost-cutting measures even before the OPEC offensive of 
 1973 quadrupled the crude oil price. Added to the ongoing dif- 
 ficulties, the OPEC price revolution broke the back of Japanese 
 industry and brought about an unprecedented two-year fall in 
 industrial output. Drastic economy measures were called for, and 
 implemented, in the use of all resources, human as well as non-human. 
 However, human resource economy posed another threat of socio- 
 political nature arising from worker resistance to employment reduc- 
 tion and livelihood deterioration. The reduction of employment and 
 
-2- 
 
 costs in Che interest of a leaner but more efficient work force there- 
 fore had to be pursued under the constraints of possible industrial 
 unrest that might be touched off by improper or sudden moves on the 
 part of the employers. A disaggregated view of how employment changed 
 in various industries and with respect to different worker and 
 employer types will be useful for shedding light on the kinds of mana- 
 gerial strategies followed for employment adjustment. As will be 
 shown below, there were predictable changes as well as a few 
 surprises. 
 
 Table 1 compares employment changes by sex and occupation between 
 1973 and 1979 across 20 branches of the manufacturing sector. It can 
 be seen that male employment decreased proportionately less than 
 female: 10 percent as against 16 percent. In addition, during the 
 same period, "managerial, clerical and technical" (non-production) 
 
 (Table 1 about here) 
 workers decreased proportionately less than "production" workers: 5 
 percent as against 15 percent. Furthermore, among the non-production 
 workers , males decreased by only two percent , while females decreased 
 by 14 percent. All this indicates that during the 1970s, Japanese 
 manufacturing work force became more masculine and more white-collar 
 oriented. But these quantitative and qualitative correlates of 
 decreasing employment are not peculiar to Japan. These are aspects of 
 commonly known economic behavior in any market economy. For example, 
 given the higher turnover of female labor than male, the use of attri- 
 tion without replacements as a means of employment reduction should 
 result in a greater reduction of female labor than male. In Japan as 
 
-3- 
 
 elsewhere, white-collar and managerial employees enjoy greater 
 employment security than production workers. This can be explained by 
 human capital theory. Although blue-collar workers are said to be 
 just as permanent in principle as white-collar workers in Japan, there 
 obviously exists a difference between principle and practice. 
 
 The Japanese hiring freeze largely falls on mid-career workers and 
 post-"teinen" workers in the process of changing jobs and young 
 workers fresh out of school looking for their first jobs. One pecu- 
 liar thing happened in the late 1970s: despite the need for work 
 force retrenchment, the "teinen" rose in many firms under pressures 
 from the government and trade unions. "Teinen" is a unique Japanese 
 institution, often incorrectly translated as "retirement," which 
 amounts to a compulsory termination of employment at a given age with 
 a severance pay amounting to roughly three and a half years of base 
 pay. The age for terminantion is conventionally 55. In recent years, 
 somewhat more than a third of major Japanese corporations still set 
 their "teinen" at this age, while a little more than a fifth of them 
 set it at 60, making the distribution of "teinen" bimodal with smaller 
 fractions of firms setting it at odd ages before 55 and between 55 and 
 60. An average would be anywhere between 55 and 60, say 56.9 years as 
 of June 1979 (Katsuyo" rodo" tokei 1982, p. 127). At the same time, 
 more consistently with the economic trends, "horror stories" about 
 earlier selective "teinen" at age 40 or 45 also abound in the press 
 with reference to major firms' economy measures. In the midst of 
 these conflicting forces, the average age and length of service in 
 Japanese manufacturing have risen as illustrated by Table 2 below. 
 
-4- 
 
 (Table 2 about here) 
 According to Table 2, the average age and the average length of 
 service rose less fast for workers with college education than for 
 those with less education. This suggests in part that the hiring 
 freeze was less severe on college graduates than on other school 
 leavers, which is consistent with the smaller reduction of employment 
 for the white-collar workers than for blue-collar as shown in Table 1. 
 An interesting point emerges from Table 2: the difference between the 
 average age and the average length of service suggests that many 
 workers must have spent varying amounts of time doing other things 
 than working for their current employers between leaving school and 
 finding jobs with the latter. (The difference between age and average 
 length of service at which a worker started to work with his current 
 employer.) For example, workers with a college education in Table 2 
 started to work for their current employers at age 25 or 26. If the 
 age of graduation from college is ideally 22, there are three or four 
 years on average unaccounted for. But here the discrepancy may not be 
 serious, because the workers in these statistics include older grad- 
 uates of prewar universities which took two more years for graduation 
 than the current system and, more importantly, because many college 
 students failed to advance to college straight from high school by 
 failing at the first entrance examinations. Considering these well- 
 known deviations from the ideal college graduation age of 22, the sta- 
 tistically inferred starting age of 25 or 26 can be considered fairly 
 consistent with the hypothesis that the jobs with their current 
 employers were the first jobs of the workers with college education. 
 
-5- 
 
 3ut the statistical starting age in Table 2 seems too high for high 
 school or junior high school leavers to support the hypothesis that 
 they were hired by their current employers fresh out of schools. The 
 normal age for leaving high school is 18, although in practice if not 
 as extensive as in the case of college students, some high school stu- 
 dents failed at the first entrance examinations and graduated later 
 than at 18. Table 2 puts the average starting age with the curent 
 employers at 23 or 24 for high school leavers, 5 or 6 years higher 
 than the high school leaving age. This suggests that many workers 
 with high school education were not hired directly from schools — an 
 important modification of the lifetime employment system. For junior 
 high school leavers, the figures are even more devastating. The 
 discrepancy between the average starting age shown in Table 2 for 
 junior high and primary school leavers, 29 or 30, and the usual school 
 leaving age of 15 must be considered a strong violation of the life- 
 time employment system. Many workers with only compulsory education 
 must have had job experiences elsewhere before they were employed by 
 their current employers. Since these workers are the mainstay of 
 Japanese blue collar, it is odd that the lifetime employment system 
 which implies employment continuity with the same employer from school 
 to teinen does not apply to them. 
 
 However, one caveat may be noted and explored. The high average 
 age of the workers with junior high school or less education may be 
 due to post-teinen workers in their late 50s or 60s who can change 
 jobs frequently. Their presence pulls up the average age but pulls 
 down the average length of service. It may be that workers typically 
 
-6- 
 
 enter permanent employment upon leaving school and stay with the same 
 employers until 55 or so, becoming mobile only after the primary life- 
 and-work cycle is completed. Data on length of service by age will 
 clarify this issue. Table 3 offers just such data. The workers in 
 the age bracket 50-54 are the most likely groups which the life-time 
 employment system would have recruited from school and kept employed 
 "for life" (until termination at 55 or so). And yet, these workers' 
 starting ages on the average are 29.7 for white collar and 35.8 for 
 blue collar. 
 
 (Table 3 about here) 
 From the above discussion of age and length of service, it is 
 clear that the meanings of "lifetime" employment and "retirement" in 
 Japan are not the same as in other countries. If "lifetime" 
 employment is only a glamorous alias for long job tenure, it is a 
 matter of degree rather than a fixed institution implying employment 
 security with the same employer over the entirety of working life from 
 school — leaving time to the time of retirement from work coinciding 
 with withdrawal from the labor force. The Japanese practice certainly 
 is not "lifetime" employment in the latter sense. In recent years, it 
 has become increasingly clear that Japanese job tenure on the whole 
 may even be inferior to American. For the U.S., Robert Hall offers a 
 convenient illustration of the proportion of long-service employees 
 (those having the same jobs for 20 or more years) by age bracket. 
 When Japanese data are put together with Hall's, Japanese job security 
 appears to be of a different kind from American. 
 
-7- 
 
 Percent of 
 Percent of workers workers with lengths 
 with same jobs for . of service for 20 or 
 Age 20 or more years, U.S. more years, Japan 
 
 7 
 
 (ID 
 
 22 
 
 (31) 
 
 23 
 
 (32) 
 
 17 
 
 (24) 
 
 13 
 
 (18) 
 
 9 
 
 (13) 
 
 11 
 
 (14) 
 
 35-39 1 
 
 40-44 7 
 
 45-49 17 
 
 50-54 25 
 
 55-59 30 
 
 60-64 33 
 
 65+ 35 
 
 For Japan, the first column refers to all employees (male and 
 female) in all industries. The second column in parenthesis refers to 
 male employees in manufacturing. It is remarkable that job tenure 
 deteriorates after age 50 in Japan, while in the U.S. it monotonically 
 improves until 65 and beyond. For people younger than 50, Japan 
 offers much more employment stability than the U.S. Japanese 
 employers have good reasons for wanting to keep younger workers in 
 their employ for as long as possible. One good reason derives 
 directly from the human capital theory of training: i.e., people are 
 more trainable and can learn faster when they are young. From an eco- 
 nomic point of view, it is an impeccable logic that the employer keeps 
 workers as long as they learn new skills and adapt to the tech- 
 nological change undertaken at the pace chosen by the employer. 
 Another side of this logic is that workers are dismissed when they age 
 and fall behind the required pace of training and learning. Japanese 
 enterprises which are known to renew their plants and capital goods 
 
-8- 
 
 much faster Chan firms in other countries have a special stake in the 
 quality of employees with reference to adaptability, trainability , and 
 versatility. When these qualities tend to diminish in a person's 
 lifetime is a question to which answers may vary. But a prudent view 
 would hold that 30-year olds can learn faster than 50-year olds. The 
 Japanese teinen for less competent employees are admirably suited to 
 Japan's national objective of rapid technological progress. 
 
 Manufacturing plants not only belong to different industries, but 
 also operate in different scales. When the employment change is 
 disaggregated on the basis of plant size, we encounter another 
 surprise: i.e., employment decreased drastically in the largest 
 plants, while it increased substantially in the smallest plants. 
 Table 4 compares employment trends in the smallest plants employing 29 
 or fewer persons with all other plants employing 30 or more persons. 
 With the 30-person plant as a divide, employment increased on the 
 smaller side but decreased on the larger side. Most of the increase 
 in employment took place in literally the smallest plants — those 
 employing nine or fewer persons. For contrast, Table 5 shows 
 employment change in the largest size class of plants employing 1,000 
 or more persons. While employment in manufacturing as a whole 
 decreased by 9 percent between 1973 and 1978, it decreased by more 
 than 20 percent in the largest plants. In the textile industry, 
 employment nearly disappeared in the largest plants, or such plants 
 nearly disappeared. Only in miscellaneous manufacturing, employment 
 increased in the largest plants. Since the largest plants are the 
 home of the Japanese employment system, the decline in the weight of 
 
-9- 
 
 these plants in the employment structure of Japanese manufacturing has 
 serious implications for the "Japanese employment system." 
 
 (Tables 4, 5, and 6 about here) 
 A finer classification of establishments and employment by size 
 presented in Table 6 confirms the role of the 30-person plant as the 
 dividing line between employment reduction and expansion. This is 
 fascinating but hard to explain. Since plants with decreasing 
 employment move down to smaller size classes when size is measured by 
 employment, it is understandable that the largest size class (1,000+) 
 lost both plants and employment between 1973 and 1978; i.e., some of 
 the plants in this class in 1973 may have been reclassified in smaller 
 size classes in 1978. For the same reason, it is no mystery that the 
 smallest size class (0-9) was the net gainer of employment, because no 
 plants could become smaller than that size class. But in the inter- 
 mediate size classes, there should have been plants joining them from 
 above by becoming smaller as well as plants moving out of them into 
 still smaller classes. The mystery is why joiners and leavers did not 
 cancel each other out. In addition, even though economic conditions 
 were generally poor during the period under consideration, there 
 should have been some growing firms. To the extent, that they crossed 
 the size boundaries upward, how the number of firms and that of their 
 employees in a given intermediate size class would change was further 
 complicated. Thus it is rather strange that a clear dividing line 
 like the 30-person plant should have emerged. Furthermore, there was 
 a net gain in the number of plants together with a net loss in total 
 employment. The new plants were in the smallest size classes, but it 
 
-10- 
 
 is not entirely clear how these small plants came into being just at 
 the time when economic difficulties were forcing the established 
 plants to cut back employment. It is possible that some of the former 
 employees of the established plants struck out on their own with or 
 without the help of their former employers. The mushrooming of small 
 entrepreneurs during a difficult period smacks of the "disguided 
 unemployment." Something to this effect is occasionally mentioned in 
 Japanese publications. However, the usual attributes of "disguised 
 unemployent , " like zero or etremely low productivity, must be rejected 
 as unreal or untrue in the case of Japanese small-scale industry, 
 because one observes a healthy growth of labor productivity in 
 
 Japanese manufacturing despite the shifting weight of employment to 
 
 3 
 smaller plants. 
 
 Japanese manufacturing is highly innovative and productive. Its 
 output, technology and productivity constitute a serious challenge to 
 world manufacturing. And yet, Japanese manufacturing is still run in 
 a style which accords a significant role to small and medium-sized 
 plants. In terms of employment, the average plant size even decreased 
 during the 1970s, from 16.9 persons per plant in 1973 to 14.6 persons 
 per plant in 1978 (Table 5). The employment share of larger plants 
 declined at the same time. One might even feel a touch of "Small is 
 Beautiful." This peculiarity of Japanese industrial structure 
 suggests that Japanese industrialization may have been supported by a 
 different combination of markets, industrial organization and tech- 
 nological choice from the Western syndrome of expanding markets, eco- 
 nomies of scale, and oligopolization. A statistical study of manu- 
 
-11- 
 
 facturing in Britain, Germany and the United States at the National 
 Institute of Economic and Social Research, England lays stress on the 
 size of plant as a major factor in labor productivity. When the 
 median size is defined as "that above (or below) which half the number 
 
 of employees in an industry are engaged," it results in 440 for Britain 
 
 4 
 (1973), 410 for Germany (1970), and 380 for the U.S. (1972). By a 
 
 similar calculation, the Japanese median plant size in 1975 comes to 
 149. When all the employed persons are taken into account, that is, 
 including not only employees but also working proprietors and their 
 working family members, the median size drops to 75. It is somewhat 
 startling that in a country where manufacturing labor productivity is 
 higher than in Britain and Germany and matches that of the U.S., a half 
 of manufacturing employment is found in plants manned by 75 or fewer 
 persons. The role of size as a source of higher labor productivity 
 via economies of scale obviously faces a Japanese challenge. 
 
 We now return to a potential puzzle implicit in the earlier data 
 (Figure 1) which showed a rise in working hours coupled with a con- 
 tinued decline in employment after 1975. The need for reducing the 
 labor input apparently ended in that year and it appears that more 
 workers could have been retained in employment if working hours had 
 not been increased. A disaggregated view of the relationship between 
 rising hours of work and declining employment is presented in Table 7. 
 
 (Table 7 about here) 
 In all two-digit industries, working hours began to increase in 1974 
 or 1975 (with one exception in 1976). But in a great majority of 
 these industries, employment continued to decrease. The question is 
 why the Japanese firms met the rising demand for labor induced by the 
 
-12- 
 
 rising level of output by reducing employment and making remaining 
 employees work, longer hours. A partial answer to that question is 
 that work sharing reducing hours of work may, beyond a point, demora- 
 lize the work force and that compared with this kind of work force, a 
 leaner work force with a full work load may produce larger and better 
 output because of improved team spirit and team efficiency. This is 
 generally known as shosu sei'ei shugi , a principle emphasizing a 
 smaller work force composed of better and abler workers. 
 Shosu sei'ei shugi would imply that the same aggregate labor hours 
 (working hours times the number of workers) would be more productive 
 when turned out by a smaller number of workers working an optimum 
 number of hours each. It is possible that at the trough of the 
 post-OPEC recession, firms had too many workers on the payroll each 
 working fewer than desirable hours. Not only in Figure 1, but also 
 on the basis of cross-industry comparison of employment and hours 
 worked, employment change is found to be negatively correlated with 
 hours worked; i.e., in industries where hours worked increased (or 
 decreased), employment decreased (or increased). Since the recovery 
 affected different industries differently, it is desirable to adjust 
 for the different rates of output expansion in order to make the 
 correlation between hours and employment purer. Thus a hypothesis to 
 be tested is: output change being the same (proxy for "other things 
 being the same"), employment should have decreased more (or less) in 
 industries in which hours worked increased more (or less). A cross- 
 industry multiple regression can then be written in the form: 
 
 (1) E = a + b H + b ? X 
 
-13- 
 
 • t 
 
 where E, H, and X are rates of change in employment, hours worked per 
 worker, and industry output. The coefficient b should be negative, 
 while b should be positive. This equation can be fitted by data in 
 Table 7 as follows: 
 
 (2) E = -2.15 - 2.04 H + 0.38 X Adjusted R = 0.3751 
 
 (1.96) (3.02) (3.21) 
 
 (Table 7 about here) 
 The fit is very good, and the coefficients for H and X in the 
 equation are highly significant (at 1 percent level) as indicated by 
 the high t-values in the parentheses. The constant term also easily 
 passes a significance test at 10 percent. The large, negative and 
 highly significant coefficient for H (-2.04) says that for every 1 
 percent increase in working hours, industries got rid of more than 2 
 percent of employees. This may imply that additional hours worked by 
 employees on the payroll are more productive than the same number of 
 hours lost by the departure of some employees. For example, in a 
 hypothetical situation where a plant employs 100 workers, each working 
 40 hours a week, additional 40 hours of work generated from 99 workers 
 (1 percent increase in aggregate hours worked) would enable the plant 
 to get rid of two workers (2 percent of employees). This should be an 
 irresistible temptation to any employer interested in raising labor 
 productivity per hour. In many manufacturing industries, working 
 hours actually averaged at about 40 hours a week in 197 5, but Japanese 
 employers and workers obviously considered that much of working hours 
 too low to be desirable. In order to ensure to each employee a 
 desirable level of working hours, then, Japanese industries had to 
 
-14- 
 
 reduce employment after 1975. One might also recall in this connec- 
 tion that American and European complaints about the "workaholic" 
 Japanese became louder during this period. Despite the rising output 
 which could have been translated into more employment, Japanese unions 
 after 1975 offered surprisingly little resistance to the on-going 
 substitution of hours for employment. Unions were satisfied if 
 appropriate conditions like generous severance pay were offered to the 
 terminated employees. In order to bring about this kind of outcome 
 favorable to themselves, employers had to keep insisting — and they 
 apparently succeeded in making everyone believe — that they were still 
 overloaded with redundant labor. 
 
 A careful econometric analysis done at the Economic Planning 
 Agency shows that excess labor supply (the difference between the 
 number of workers firms had in their employ and the number they con- 
 sidered justified by the level of output and that of regular working 
 hours) had disappeared in Japanese manufacturing by mid-1975. This 
 analysis treats all the working hours as homogeneous, whereas Equation 
 2 above suggests that in and after 1975, additional hours worked by 
 remaining employees were more productive than the same number of hours 
 that the departing employees would have supplied. It appears, then, 
 that the total number of hours from 100 workers at X hours each (100X) 
 is less productive than the same number of hours worked by 80 workers 
 at 5/4 of X hours each (still 100X). Why a leaner work force can be 
 Qore efficient than a fatter work force for the same aggregate number 
 of hours cannot be explained without discarding the assumption that an 
 hour of work is an hour of work regardless of who puts it in or when. 
 A morale factor which can be generated more easily in a leaner work 
 
-15- 
 
 force can be "labor-augmenting" in the sense that an hour of work com- 
 bined with high morale is more productive than an hour of demoralized 
 work. Japanese employers obviously took advantage of the post-OPEC 
 economic crisis for inspiring workers with the need for renewed dedi- 
 cation and discipline. Government helped them in this campaign. 
 Those workers who had to lose jobs or forego opportunities in manufac- 
 turing probably did so in the belief that their loss of jobs or oppor- 
 tunities was in the interest of more efficient Japanese industry. 
 
 Conclusion 
 
 There is something extraordinary in the concentration of Japanese 
 efforts to improve labor productivity in manufacturing with heavy em- 
 phasis on the elimination of employment, apparently unmindful of pro- 
 ductivity in all other sectors. What labor shedding in manufacturing 
 amounts to , then, is a shift of the burden of employment creation to 
 other sectors. As a consequence these sectors suffer from downward 
 pressure on their productivity growth. This conceivable negative inter- 
 dependence between manufacturing and non-manufacturing has so far es- 
 caped observers' serious attention. But there is no doubt that it is 
 a subject requiring urgent and careful analysis. Aware of this need, 
 this paper examines the nature of employment adjustment in manufactur- 
 ing and suggests corrections with respect to what must now be consid- 
 ered misconceptions of Japanese employment practices like lifetime 
 employment. 
 
-16- 
 
 Footnotes 
 
 Robert E. Hall, "The Importance of Lifetime Jobs in the U.S. 
 Economy," American Economic Review , 72, 4 (September 1982): 717. 
 The figures are based on the Current Population Survey . "A job is 
 defined as continuous employment with the same employer, possibly in 
 
 different occupations," the author says. 
 
 2 
 Calculated from Ministry of Labor, Yearbook of Labor Statistics , 
 
 1979, pp. 176-192. 
 
 3 
 An illuminating discussion of size differences in employment 
 
 adjustment is Eiko Shinotsuka, "Kigyo kibobetsu ni mita saikin no koyo 
 
 chosei" (Employment Adjustment by Size of Enterprise), Nihon Rodo 
 
 Kyokai Zasshi (The Journal of the Japan Institute of Labor) (February 
 
 1979): 2-13. 
 
 4 
 S. J. Prais et al., Productivity and Industrial Structure (London: 
 
 Cambridge University Press, 1981), pp. 9-10. 
 
 Since the OPEC shock, there has been a sort of fixation on the 
 central importance of the labor costs among Japanese enterprises, as 
 discovered by a survey undertaken by MITI. See MITI, Nihonteki koyo 
 kanko no yuku'e (Whither Japanese Employment Institutions?) (Tokyo: 
 Sangyo Noritsu Daigaku Press, 1981). 
 
 Haruo Shimade et al. , Rodo shijo kiko no kenkyu (Studies in the 
 Labor Market Mechanism) (EPA Research Series No. 37) (Tokyo, 1981), 
 pp. 311-322, 384-395. 
 
 When an employee's efficiency per day is a nonlinear function of 
 the hours of work, the profit-maximizing employer must determine the 
 optimum number of hours each worker should work per day and the opti- 
 
-17- 
 
 mum number of workers. That is, hours per worker and employment 
 become two independent variables. The analytical problem here is for- 
 mally similar to the theory of underemployment presented by Harvey 
 Leibenstein a generation ago: Harvey Leibenstein, Economic Backward- 
 ness and Economic Growth (New York: Wiley, 1957), Chapter 6. For a 
 further exploration of puzzles in the determination of optimum hours 
 and optimum employment by a profit-maximizing entrepreneur, see Koji 
 Taira, "Surplus Labor, Undernourishment, and Labor Market Rationality 
 in LDC Agriculture: The Leibenstein Model," Keizai Kenkyu (Economic 
 Review), 21, 4 (October 1970): 312-322. 
 
 D/121 
 
I J 
 
 -j- 
 
 r r 1 
 
 jFiguiS 1. r 
 
 : hi" 
 
 ! 1 
 
 ■ft 
 
 . 4 — L 
 
 ! 
 
 I 
 
 f 
 
 ■4— \. 
 
 i 
 
 ■ i 
 
 i 
 
 
 • -< 
 
 H- j - i 
 
 .J- J 
 
 ..... ; 
 
 I 
 
 4 -J- 
 
 i ;-i 
 
 1 i 
 
 t""T Pti 
 
 Hours: worked..pex month pejr worker and i}he: index of employment qf 
 
 1. 
 
 regular workers in manufacturing establishments employing 30 i 
 
 : i t : ~j — ! — | ! fTT l r r [" j i 1" I 
 
 more regular w or ker s ('index: i980 — 100) |— j 
 
 or 
 
 ! ; 
 
 ■■1 r 
 -t~t—t 
 
 ...i 
 
 t- 
 
 1 1.+4444..4 
 
 44-4 M-H^M 
 
 ' -4- ! -I -I 
 
 I - -• i-1 
 
 , -4 — t — t — t i- j-t--t- 
 
 -_ J— L 
 
 • r 
 
 i:m 
 
 *. 
 
 i- 
 
 1 -; 
 
 7r 
 
 ! 1 
 
 I I 
 
 — - -1 
 
 4- 
 
 i._ 
 
 -f-4- 
 
 tto 
 
 I 
 
 j«i 
 
 if»i 
 
 r - 
 
 f 
 
 i 
 
 _ L 
 
 -f -'- -4 t- 
 
 — i - ■ ■+ • 
 
 - i I i- ; 
 
 i- - 
 
 I i Tj 
 
 1 . — ! i 
 
 1-'T 
 
Table 1. Ratio of Employment in 1979 to that in 1973 by 
 
 Industry, Occupation and Sex: Japanese Manufacturing 
 
 Managerial, clerical 
 All Workers Production Workers and technical worker 
 
 Industry 
 
 Total 
 
 Male 
 
 Female 
 
 Total 
 
 Male 
 
 Female 
 
 Total 
 
 Male 
 
 Female 
 
 Total 
 
 .88 
 
 .90 
 
 .84 
 
 .85 
 
 .86 
 
 .84 
 
 .95 
 
 .98 
 
 .86 
 
 18/19 
 
 1.11 
 
 1.13 
 
 1.08 
 
 1.09 
 
 1.11 
 
 1.01 
 
 1.15 
 
 1.18 
 
 1.09 
 
 20 
 
 .65 
 
 .74 
 
 .59 
 
 .62 
 
 .71 
 
 .57 
 
 .79 
 
 .80 
 
 .78 
 
 21 
 
 1.11 
 
 1.13 
 
 1.10 
 
 1.12 
 
 1.12 
 
 1.12 
 
 1.06 
 
 1.14 
 
 1.00 
 
 22 
 
 .69 
 
 .73 
 
 .61 
 
 .66 
 
 .71 
 
 .57 
 
 .82 
 
 .81 
 
 .90 
 
 23 
 
 .89 
 
 .90 
 
 .87 
 
 .85 
 
 .86 
 
 .86 
 
 1.00 
 
 1.00 
 
 1.00 
 
 24 
 
 .91 
 
 .97 
 
 .75 
 
 .90 
 
 .97 
 
 .72 
 
 .93 
 
 .95 
 
 .93 
 
 25 
 
 .94 
 
 .96 
 
 .85 
 
 .95 
 
 .97 
 
 .88 
 
 .94 
 
 .97 
 
 .85 
 
 26 
 
 .86 
 
 .92 
 
 .70 
 
 .82 
 
 .89 
 
 .65 
 
 .89 
 
 .95 
 
 .74 
 
 27 
 
 .76 
 
 .91 
 
 .86 
 
 .96 
 
 .96 
 
 1.00 
 
 .88 
 
 .89 
 
 1.00 
 
 28 
 
 .81 
 
 .87 
 
 .71 
 
 .81 
 
 .88 
 
 .70 
 
 .81 
 
 .87 
 
 .80 
 
 29 
 
 1.03 
 
 .94 
 
 1.23 
 
 1.00 
 
 .92 
 
 1.27 
 
 1.40 
 
 1.25 
 
 2.00 
 
 30 
 
 .91 
 
 .94 
 
 .84 
 
 .87 
 
 .90 
 
 .79 
 
 1.07 
 
 1.08 
 
 1.05 
 
 31 
 
 .84 
 
 .85 
 
 .71 
 
 .83 
 
 .85 
 
 .63 
 
 .85 
 
 .88 
 
 .77 
 
 32 
 
 .87 
 
 .91 
 
 .74 
 
 .88 
 
 .90 
 
 .74 
 
 .87 
 
 .93 
 
 .73 
 
 33 
 
 .82 
 
 .83 
 
 .83 
 
 .81 
 
 .1 
 
 .82 
 
 .86 
 
 .88 
 
 .84 
 
 34 
 
 .83 
 
 .83 
 
 .82 
 
 .77 
 
 .77 
 
 .78 
 
 .93 
 
 .96 
 
 .86 
 
 35 
 
 .93 
 
 .99 
 
 .85 
 
 .90 
 
 .96 
 
 .85 
 
 .98 
 
 1.02 
 
 .83 
 
 36 
 
 .84 
 
 .84 
 
 .83 
 
 .80 
 
 .80 
 
 .88 
 
 .93 
 
 .97 
 
 .80 
 
 37 
 
 1.02 
 
 1.04 
 
 1.00 
 
 .99 
 
 .97 
 
 1.00 
 
 1.12 
 
 1.16 
 
 1.00 
 
 38/39 
 
 1.03 
 
 1.02 
 
 1.06 
 
 .98 
 
 .92 
 
 1.08 
 
 1.17 
 
 1.22 
 
 1.04 
 
 Sources: Ministry of Labor, Year Book of Labour Statistics, 1973, 1979. 
 
Notes Co Table 1 
 
 18/19: Food and kindred products, tobacco manufacturing; 
 
 20: Textile mill products; 
 
 21: Apparel and other finished products made from fabrics and similar 
 materials; 
 
 22: Lumber and wood products; 
 
 23: Furniture and fixtures; 
 
 24: Pulp, paper and paper worked products; 
 
 25: Publishing, printing and allied industries; 
 
 26: Chemical and allied products; 
 
 27: Petroleum and coal products; 
 
 28: Rubber products; 
 
 29: Leather and leather products; 
 
 30: Ceramic, stone and clay products; 
 
 31: Iron and steel; 
 
 32: Non-ferrous metal products; 
 
 33: Fabricated and metal products; 
 
 34: Machinery; 
 
 35: Electrical machinery; 
 
 36: Transportation and equipment; 
 
 37: Precision machinery; 
 
 38/39: Miscellaneous manufacturing industries. 
 
Table 2: The Average and Length of Service For Male Employees 
 in Japanese Industry by Level of Education, 1973 and 
 1979. (in years) 
 
 
 All Industries 
 
 Manufacturing 
 
 
 1973 
 
 1979 
 
 Change 
 
 1973 
 
 1979 
 
 Change 
 
 University graduates 
 
 
 
 
 
 
 
 Age 
 
 34.0 
 
 34.4 
 
 -K).4 
 
 33.7 
 
 35.0 
 
 +1.3 
 
 Length of service 
 
 8.4 
 
 8.9 
 
 -KD. 5 
 
 8.5 
 
 9.9 
 
 +1.4 
 
 Difference 
 
 25.6 
 
 25.5 
 
 
 25.2 
 
 25.1 
 
 
 High school graduates 
 
 Age 32.4 34.7 +2.3 31.6 34.3 +2.7 
 
 Length of service 8.5 9.9 +1.4 8.5 10.8 +2.3 
 
 Difference 23.9 24.8 23.1 23.5 
 
 Junior high and primary 
 school leavers 
 
 Age 
 
 39.8 
 
 42.5 
 
 +2.7 
 
 38.9 
 
 41.8 
 
 +2.9 
 
 Length of service 
 
 10.6 
 
 11.6 
 
 +1.0 
 
 11.1 
 
 13.2 
 
 +2.1 
 
 Difference 
 
 29.2 
 
 30.9 
 
 
 27.8 
 
 28.6 
 
 
 Sources : Japan Productivity Center, Katsuyo rodo tokei (Practical Labor 
 Statistics), 1975 and 1981. _The original data_sources are 
 Ministry of Labor, Chingin kozo kihon tokei chosa (popularly 
 known as "Wage Census"). 
 
Table 3. Average ages and lengths of service of male manufacturing 
 workers by occupation, 1979 
 
 
 All 
 
 workers 
 
 White 
 
 collar 
 
 Blue 
 
 collar 
 
 
 
 Average 
 
 
 Average 
 
 
 Average 
 
 
 Average 
 
 length of 
 
 Average 
 
 length of 
 
 Average 
 
 length of 
 
 Age 
 
 age 
 
 service 
 
 age 
 
 service 
 
 age 
 
 age 
 
 All ages 
 
 37.6 
 
 11.7 
 
 37.7 
 
 13.0 
 
 37.5 
 
 10.9 
 
 30-34 
 
 32.3 
 
 10.0 
 
 32.5 
 
 9.9 
 
 32.3 
 
 10.2 
 
 35-39 
 
 37.4 
 
 10.0 
 
 37.4 
 
 14.3 
 
 37.5 
 
 12.5 
 
 40-44 
 
 42.5 
 
 15.2 
 
 42.4 
 
 17.4 
 
 42.5 
 
 13.8 
 
 45-49 
 
 47.4 
 
 17.1 
 
 47.3 
 
 20.4 
 
 47.4 
 
 15.3 
 
 50-54 
 
 52.4 
 
 18.7 
 
 52.4 
 
 22.7 
 
 52.3 
 
 16.6 
 
 55-59 
 
 57.2 
 
 15.7 
 
 57.1 
 
 19.0 
 
 57.2 
 
 14.1 
 
 60-64 
 
 62.3 
 
 11.2 
 
 62.3 
 
 12.8 
 
 62.3 
 
 10.5 
 
 65+ 
 
 68.8 
 
 12.3 
 
 69.0 
 
 13.9 
 
 68.7 
 
 11.6 
 
 Sources: Ministry of Labor, Yearbook of Labor Statistics , 1979, 
 
 pp. 162-164. The original sources are the "Wage Censuses.' 
 
Table 4. Employed Persons in Manufacturing Size of Establishment 
 1970-1978: Absolute Numbers in Indices 
 
 Year Total 
 
 Establishments 
 employing 0-9 
 
 sersons 
 
 Establishments 
 
 employing 10-19 
 
 persons 
 
 Establishments 
 employing 20-29 
 persons 
 
 Establishments 
 employing 30+ 
 persons 
 
 (Absolute number , in thousands) 
 
 1970 
 
 11,679 
 
 1,910 
 
 1,262 
 
 647 
 
 
 7,862 
 
 1971 
 
 11,463 
 
 1,886 
 
 1,216 
 
 636 
 
 
 7,726 
 
 1972 
 
 11,783 
 
 2,056 
 
 1,315 
 
 669 
 
 
 7,743 
 
 1973 
 
 11,961 
 
 2,079 
 
 1,325 
 
 683 
 
 
 7,874 
 
 1974 
 
 11,486 
 
 2,035 
 
 1,255 
 
 675 
 
 
 7,521 
 
 1975 
 
 11,296 
 
 2,152 
 
 1,283 
 
 690 
 
 
 7,171 
 
 1976 
 
 11,173 
 
 2,112 
 
 1,158 
 
 840 
 
 
 7,064 
 
 1977 
 
 10,874 
 
 2,078 
 
 1,120 
 
 848 
 
 
 6,828 
 
 1978 
 
 10,890 
 
 2,178 
 
 1,151 
 
 886 
 
 
 6,675 
 
 
 
 
 (Index, 1970 = 100) 
 
 
 
 1970 
 
 100 
 
 100 
 
 100 
 
 100 
 
 
 100 
 
 1971 
 
 98.2 
 
 98.7 
 
 96.4 
 
 98. 
 
 3 
 
 98.3 
 
 1972 
 
 100.9 
 
 107.6 
 
 104.2 
 
 103. 
 
 4 
 
 98.5 
 
 1973 
 
 102.4 
 
 108.8 
 
 105.0 
 
 105. 
 
 6 
 
 100.2 
 
 1974 
 
 98.3 
 
 106.5 
 
 100.3 
 
 104. 
 
 3 
 
 95.7 
 
 1975 
 
 96.7 
 
 112.7 
 
 101.7 
 
 106. 
 
 6 
 
 91.2 
 
 1976 
 
 95.7 
 
 110.6 
 
 91.8 
 
 129. 
 
 8 
 
 89.9 
 
 1977 
 
 93.1 
 
 108.8 
 
 88.7 
 
 131. 
 
 ,1 
 
 86.9 
 
 1978 
 
 93.2 
 
 114.0 
 
 91.2 
 
 136. 
 
 9 
 
 84.9 
 
 Sources: MIT I , Census of Manfacturers. 
 
Table 5. Employment in Manufacturing As A Whole and In Establishments 
 
 Employing 1,000 Persons or More, by Industry, in 1973 and 1978 
 
 (in 000s of persons) 
 
 Establishments employing 
 Manufacturing as a Whole 1,000 or more persons 
 
 1973 1978 '78/73 
 
 Industrv 
 
 1973 
 
 1978 
 
 *78/*73 
 
 Total 
 
 11,961 
 
 10,890 
 
 0.91 
 
 18/19 
 
 1,153 
 
 1,165 
 
 1.01 
 
 20 
 
 1,167 
 
 867 
 
 0.74 
 
 21 
 
 497 
 
 547 
 
 1.10 
 
 22 
 
 510 
 
 415 
 
 0.81 
 
 23 
 
 317 
 
 306 
 
 0.96 
 
 24 
 
 333 
 
 300 
 
 0.90 
 
 25 
 
 486 
 
 505 
 
 1.04 
 
 26 
 
 466 
 
 415 
 
 0.89 
 
 27 
 
 44 
 
 45 
 
 1.02 
 
 28 
 
 170 
 
 157 
 
 0.92 
 
 29 
 
 86 
 
 94 
 
 1.09 
 
 30 
 
 585 
 
 533 
 
 0.91 
 
 31 
 
 528 
 
 449 
 
 0.85 
 
 32 
 
 228 
 
 191 
 
 0.84 
 
 33 
 
 910 
 
 829 
 
 0.91 
 
 34 
 
 1,193 
 
 1,054 
 
 0.88 
 
 35 
 
 1,399 
 
 1,240 
 
 0.89 
 
 36 
 
 985 
 
 893 
 
 0.91 
 
 37 
 
 261 
 
 258 
 
 0.99 
 
 38 
 
 2 
 
 1.3 
 
 0.65 
 
 39 
 
 642 
 
 627 
 
 0.98 
 
 1,935 1,512 0.78 
 
 19 16 0.84 
 
 58 4 0.07 
 
 5 x 
 
 x 8 
 
 9 6 0.67 
 
 23 21 0.91 
 
 57 43 0.75 
 
 141 102 0.72 
 
 x x 
 
 49 36 0.73 
 x 2 
 
 34 23 0.68 
 
 235 207 0.88 
 
 62 42 0.68 
 
 17 9 0.53 
 
 234 202 0.86 
 
 433 341 0.79 
 
 495 403 0.81 
 
 50 34 0.68 
 x x 
 
 17 22 1.29 
 
 Sources: MITI, Census of Manufactures , 1973 and 1978. "x" indicates that 
 there are too few establishments (one or two) to protect them 
 against identification if the required figures are published. 
 
Table 6: The Number of Establishments and the 
 by Size of Establishment in Japanese 
 
 Number of Employed Persons 
 Manufacturing, 1973 and 197! 
 
 Size class by 
 number of employed 
 persons 
 
 Establishments 
 
 (in number of units) 
 
 1973 1978 Change 
 
 527,470 569,866 +42,396 
 
 Employment 
 (in 000s of perso 
 1973 1978 Ch 
 
 2,079 2,178 + 
 
 ns) 
 
 ange 
 
 9 or fewer 
 
 99 
 
 10-19 
 
 93,461 
 
 83,689 
 
 - 9,772 
 
 1,325 
 
 1,152 
 
 - 
 
 173 
 
 20-29 
 
 27,763 
 
 36,629 
 
 + 8,866 
 
 683 
 
 885 
 
 + 
 
 202 
 
 30-49 
 
 25,376 
 
 22,265 
 
 - 3,111 
 
 981 
 
 856 
 
 - 
 
 125 
 
 50-99 
 
 18,302 
 
 18,164 
 
 138 
 
 1,277 
 
 1,245 
 
 - 
 
 32 
 
 100-199 
 
 8,959 
 
 7,905 
 
 - 1,054 
 
 1,231 
 
 1,085 
 
 - 
 
 146 
 
 200-299 
 
 2,796 
 
 2,326 
 
 470 
 
 677 
 
 566 
 
 - 
 
 111 
 
 300-499 
 
 2,043 
 
 1,694 
 
 349 
 
 781 
 
 650 
 
 - 
 
 131 
 
 500-999 
 
 1,433 
 
 1,126 
 
 307 
 
 991 
 
 762 
 
 - 
 
 229 
 
 1,000+ 
 
 844 
 
 673 
 
 171 
 
 1,935 
 
 1,512 
 
 - 
 
 423 
 
 Total 
 
 708,447 
 
 744,337 
 
 +35,890 
 
 11,961 
 
 10,890 
 
 -1 
 
 ,071 
 
 Per establishment 
 
 _ 
 
 _ 
 
 — 
 
 16.9 
 
 14.6 
 
 
 
 Sources: MITI, Census of Manufactures, 1973 and 1978. 
 
Table 7. The Rates of Change in Employment, Hours Worked and Output in 
 Manufacturing Industries, from the Trough of the Recession to 
 1978. 
 
 
 Average annual 
 
 Average annual 
 
 Average annual 
 
 
 rate 
 
 of change 
 
 rate 
 
 of change 
 
 rate of change 
 
 Industry 
 
 in 
 
 employment 
 
 in hours worked 
 
 in output , 
 
 code-*- 
 
 1975 to 1979 2 
 
 1975 
 
 to 1979 2 
 
 1975 to 19783 
 
 
 
 (%) 
 
 
 (%) 
 
 <%) 
 
 18/19 
 
 
 2.03 
 
 
 .28 
 
 4.4 
 
 20 
 
 
 -5.16 
 
 
 1.02 
 
 3.3 
 
 21 
 
 
 .52 
 
 
 .69 
 
 10.2 
 
 22 
 
 
 -5.75 
 
 
 2.12 
 
 6.4 
 
 23 
 
 
 - .63 
 
 
 1.19 
 
 8.3 
 
 24 
 
 
 -2.06 
 
 
 1.77 
 
 8.4 
 
 25 
 
 
 -1.40 
 
 
 .70 
 
 8.6 
 
 26 
 
 
 -2.52 
 
 
 .77 
 
 11.0 
 
 27 
 
 
 -1.93 
 
 
 .06 
 
 1.0 
 
 28 
 
 
 -1.85 
 
 
 1.73 
 
 9.0 
 
 29 
 
 
 - .15 
 
 
 .48 
 
 10.0 
 
 30 
 
 
 -1.40 
 
 
 1.37 
 
 6.8 
 
 31 
 
 
 -4.58 
 
 
 1.38 
 
 8.8 
 
 32 
 
 
 -1.93 
 
 
 2.45 
 
 16.8 
 
 33 
 
 
 -2.06 
 
 
 2.21 
 
 7.0 
 
 34 
 
 
 -3.03 
 
 
 2.16 
 
 6.8 
 
 35 
 
 
 1.35 
 
 
 2.11 
 
 18.0 
 
 36 
 
 
 -3.73 
 
 
 1.64 
 
 8.5 
 
 37 
 
 
 1.85 
 
 
 1.46 
 
 13.2 
 
 38/39 
 
 
 2.15 
 
 
 .91 
 
 10.0 
 
 Notes to Table 7 
 
 See Table 1. 
 
 "The negative rates of increase are calculated as follows: 
 (The value of X in 1979) - (The value of X in 1975) x 
 (1-r)"* where r is the rate of decrease. 
 
 The positive rates of increase are calculated as follows: 
 (The value of X in 1979) = (The value of X in 1975) x 
 (1+R)^ where R is the rate of increase. 
 
 Sources: Ministry of Labor, Yearbook of Labor Statistics , 1975 
 and 1979. The data refer to manufacturing establishments 
 employing 30 or more regular workers. 
 
 Calculated from data on values added in various industries 
 deflated by wholesale price indices specific to their products. 
 Sources: MITI, Census of Manufactures, 1978. 
 
HECKMAN 
 BINDERY INC. |S 
 
 JUN95 
 
 „ , t p, J N. MANCHESTER.