firm bize differentiation in japan - mcgill...
TRANSCRIPT
1
FIRM BIZE DIFFERENTIATION IN JAPAN
Themistoklis Chrisanthopoulos
Department of Sociology
McGill University, Montreal
September 1991
A Thesis subrni tted to the Facul ty of Graduate Studies and
Research in partial fulfillment of the requirements of the
degree of Master of Arts.
(c) Themistoklis Chrisanthopoulos, 1991
J
. 1
ABSTRACT
The Japanese labour market is analyzed wi thin a dual
labour market context and an emphasis on firm size
differentiation. Labour market segmentation theories are
presented for the purpose of undrrstancUn7 the differences
that can exist between industries, or in this case, between
firrn sizes. Data on labo~~ market variables such as wages,
unionism, and promotion illustrates the different employment
package that employees face in small firms as compared to
workers in large companies. Unique Japanese labour market
traits such as the expectation of lifeti"tle employment,
enterprise unionism, and seniority-based wages are also
observed to evaluate their application in different company
si zes. Japan' s labour market segmentation is pr imar i ly a
study of differences between firm sizes. The dissimilarities
in employment characteristics between small and large firms
are thoroughly examined and evaluated for a solid
understanding of Japan's duality in the labour market .
i
( 'J
RESUME
Le marché du travail japonais est analysé dans un
contexte de dualité avec une attention particulière portée sur
la différentiation des t ... d Iles des firmes. Af in de comprendre
les différences qui peuvent exister entre les industries, ou
dans le cas présente lçi, entre les tailles des firmes, des
théories sur la segmentation du marché du travail seront
présentées. PlusieurF; variables dans le marché du travail
telles les salaires, le taux de syndicalisation et les
politiques de promotion serv'nt à illustrer les différentes
conditions de travail vécues par des (:!mployés de peti tes
firmes relativement aux employés de grandes firmes. certains
traits sont uniques au marché du travail japonais tels
l'espérance d'un emploi à vie, la syndicalisation d'entreprise
et les salaires basés sur l'ancienneté; ces car3ctéristiques
sont observées afl n d'évaluer leur mise-en-place dans des
firmes de tailles différentes. La segmentation du marché du
travail au Japon est essentiellement l'étude des différences
entre les entreprises de tailles différents. Les différences
des caractéristiques d'emploi entre les petites et grandes
firmes seront rigoureusement étudiées et évaluées afin d~en
arriver à une compréhension soiide de la dualité du marché du
travail au Japon.
J ii
1
ACRNOW~,EDGEMENT
l thank Mr. Paul T. Dickinson for pr;:>viding heIpful
comments and encouraging my pursuit to study the Japanese
labour market. ProfesE::or Michael R. smith also provided
constructi ve ideas in the preparation of thü.~ thesis. l
deeply appreciate the continuaI support and understanding of
my famiIy and friends.
iii
To my parents,
iv
J
1
1
TABLE OF CONTENTS
ABSTRACT.
RESUME. .
ACKNOWLEDGEMENT.
LIST OF TABLES.
INTRODUCTION
CHAPT ER ONE: LABOUR MARKE']' SEGMENTATION THEORY
1. Origins of Labour Market Segmentation
2. Reasons for Labour Market Se9mentation
3. Institutionalist Theory
a) InternaI Labour Markets . b) Neoinstitutionalists . . ..• c) Human Capital Theory
. .
Page
. . . i
ii
~ii
. . vii
1
8
9
10
12
13 15 16
4. Historical ~rocess of the American Labour Market. 16
5. Dual Labour Market Theory primary vs. Secondary Labour Markets . . • 18
6. Upper-tier and Lower-tier vs. Secondary 20
7. Labour Market Segmentation . 22
8. Evolution of Dualism 28
9. Support for Duali ty 29
10. criticisms of Dual Labour Market Theory 30
CHAPTER TWO: 'l'HE THREE PILLARS AND WAGES 32
1. Japan as an Economie Suc cess Story • 32
The Three Pillars 34
2. Lifetime Employment 34
v
l
3. Enterprise Unionism
4. Seniority-based Wage System
Wage Data
5. dlue-collar Workers . . . . 6. White-collar Workers .
7. Male Earnings - a Il Industries .
8. Female Earnings - all Industries.
9 < Wage Setting (Shunto)
CHAPTER THREE: LARGE VS. SMALL FIRMS
1. Women in the Japanese Labour Market
2. Education
Large vs. Small Firms
:; . InternaI Promotion . .
4. Internal Training and Mobility
5. Job Tenure and 1eparation Rates
6. Workweek dnd Overtime
7. Quality Control Circles and Joint Consultation . . . . .
8. Proctvçtivity DifferentiaIs.
CONCLUSION
BIBLIOGRAPH'i. .
vi
Page
38
42
42
43
45
46
49
52
55
55
59
62
64
66
67
69
70
72
73
. . . . 80
1
2.1
2.2
3.1
3.2
3.3
,
LIST OF TABLES
Monthly Male Earnings by Firrn Size for 20-24 Age Group
Fernale Monthly Contractual Earnings
Starting Wages in Sakai
White-Collar Female Wages by Firm Size & Blue-Collar Fernale Wages by Firm Size
Average Annual Separation Rate by Firrn Size
vii
47
49
57
58
68
J
INTRODUCTIO~
The Japanese labour-management framework will he ana 1 yzed
in order ta assess the extent of labour market segmentation,
'J'lith an emphasis on finn Slze differentiatlon, within the
Japanese labour market. There has been an enormous dmount of
research on labour market segmentation in the UnIted stotes.
Many economists and soc iolog ists who hav(~ tack l cd Amcr! Cdn
labour market segmentation have concluded that segmentat i on in
employment benef i ts occurs predomi nant ly betwcen i ndustr i es 1 n
the v.s. (Piore, 1975; Averitt, 1968). Research 1S lacklng,
however, in determining the form of labour market segmentat Ion
that exists in Japan and if V:tbour market dual ity theories are
relevant in descr ibing the Japanese l.1bour market. Even
though Japan's economic progress has been phenomenal ln the
last few decades, its form of labour market segmentation has
rarely been observed. l nstead, most of the resea rch on the
Japanesp. labour market eleals with Japan's unique management or
employee characteristics, Japanese firm ideology such as ~ork
ethic and cultural traits. 'J'he widely used labour market
segmentation the ory which attempts to describe the labour
market in the United states, has not been utilized to describe
Japan's labour market. My thesis will focus on testing the
applicability of firm size differentiation arguments, found in
t dual labour market theory, in Japan.
2
The an,'\lysis will
consist of severa l chapters in order ta present the labour
market segmentation theory, Japan's labour market
charélcter ist ics, and the appl icabi] i ty of a dual labour market
in Japan.
In the f irst chapter, an examination af the labour market
segmentation theory wj Il be presented in order to later
evaluate the existence and type of labour li\arket segmentation
in Japan. Hodson and Kaufman's (1982) account of labour
market dua li ty wi Il be extensi vely presented ta explain the
basic foundation of the theory and l ts cri t icisms. Gordon
(1982), Pion~ (lQ75), Averitt (1968), and other mainstream
labour markt.,t segmentation advocates will complete the
analysis of the theoretical assertions. Labour market
segmentation' s pivotaI point of firm size diff,erentiation will
be the primary concern. For a labour market to be considered
dual, accarding ta firm size, a primary sector which consists
of large enterprises and a secondary sector mainly of small
companies must coexist. The two different sizes of firrns,
large and small, demonstrate significant opposing
character istics. The large enterprises of the core sector are
charactel i zed by h igh wages, relati vely good employment
benefi ts, and internaI labour market rnechanisms such as
internaI promotion schemes. There is also a higher degree of
unionization and internaI training in large firms. Small
companies in the peripheral sector of the economy are
J
3
characterized by low wages, external operating schemes, few
employment benefi ts, and limited opportuni ties for
advêtncement. Unionization is rare and ext.ernal training is
usually required. The first chapter wi] l examine these
opposing labour market traits as weIl as other aspects of
labour market segmentation theory.
since dual labour market ideology originated and
continued to develop predominantly in the United states, and
not Japan, the literature used in the first chapter to explain
the var ious f acets of labot,;r market segmentat ion will be
American. In essence, an American labour market theoretical
framework will he analyzed in arder to evaluate its
applicabili ty ta another labour market. - Japan. Aver i t t' s
(1968) analysis will be provided as one of :'he original
research works on the topic of a dual labour market theory.
His analysis focused on the dual nature of the economy as a
whole. Averi tt presented a clear distinction bet ween the
"core" and the "periphery" sectors of the economy. He clearly
distinguished between large firms in the core sector and small
firrns in the periphery. Averitt argued that large iirms use
advanceà techn/..,log.lcal equipment and are diversified into many
industries, regions, and nations. Small firms, on the other
hand, are usually owned and managed by an indi v idua] or a
family and are invol ved in local markets. The distinction
between large and smûll firms, made by Averitt in 1968, was
one of the first discussions of the existence of duality in
" !
4
the labour market.
Since Averi tt' s analysis, many labour market segmentation
advocatcs have followed. Piore (1975), for instance, provides
a description of the pr imary and secondary labour markets
whereby large firms, in the primary sector, consist of upper-
tier and lower-tier occupations. Piore indicates that there
are three segments in labour market analysis. There are large
f irms wi th a separa te upper and lower-tier versus small
companies. The upper-tier of large enterprises mainly
contains professional and managerial positions. The lower-
tier consists primar1ly of production workers. Moreover,
Gordon' s (1982) analysis on the evolution of the labour
movement into a dual existence in the United states is helpful
in understanding the process which lead to labour market
segmentation. Gordon presents three periods between the
1820's and the present in which the labour movement followed
a path whereby labour market segmentation was the outcome.
Thus, a brief historical development of the dual labour market
the ory by Gordon (1982) will provide sorne understanding into
the present arguments of labour market segmentation. Hodson
and Kaufman' s (1982) contribution to labour market
segmentation will provide a final thorough examination of the
theory's main emphasis of firm size differentiation within two
distinct labour markets.
The following two chapt ers will provide a description of
certain Japanese labour market characteristics with a focus on
5
differences depending on company size. In Chapter Two, the
three pillars of the Japanese labour market will be presented.
Dore (1989), Koike (1988), nnd Weiler (lj86) will contribute
the main data on lifetime employment, enterprise unionism, and
the seniority-based w~ge system. These three pillaLs will be
discussed in order to observe where the y are operatjng. Are
they found within large or small firms? Furthermore, wage
data will be presented. Mitsuru's (1990) and Koike's (1988)
wage results on blue-caIlar workers, white-caIlar employees,
female and male earnings will be provided. The findings on
wages will determine where the seniority-based wage system is
found and will present the wage differentials that exist
between large and small companies. After examining the three
pillars and wages in Japan, the applicability of labour mark.et
segmentation by firm size will be evaluat~d.
In Chapter Three, other Japanese labour market variables
will be examined in order ta assess, once again, the relevance
of firm size differentiation as the contributing factor to
Japan's labour market segmentation.
characteristics in comparison ta
First, women's employment
men' s will be provided.
Women's labour market traits, such as wages, will be examined
according ta company size. Japan~~se female labour market data
will be prirnarily taken from Mitsuru (1990) and Koike (1988).
Secondly, education in Japan will be observed in order ta
assess the hiring and promotion practices of large and small
firms. Finally, an analysis of large and sr.all corn(: 3nies wi Il
attempt to determine if differences existe
6
Labour market
traits su ch as internaI promotion, training, and mobility,
will be presented. Data on job tenure, separation rates,
workweek patterns, overtime, quali ty control circles, and
joint consultation sessions will supply information for an
assessment of Japan's labour market segmentation contingent
upon the size of a company. Data will be gathered
predominantly by Mitsuru (1990), Dore (1989), Koike (1988),
and Weiler (1986).
A conclusion will follow with arguments on Japan's
dualism by Kalleberg (1988) and Taira (1985). Moreover, the
results of the previous chapters will be completely examined
for a final statement on Japan' s form of labour r'.arket
segmentation. The entire emphasis will be on observing if the
Japanese labour market data demonstrates that firm size
differentiation is the pivotaI factor in determining a
worker's employment traits. After presenting Japanese labour
market characteristics, l will he able to observe if large
enterprises follow the description of labour market
segmentation views. Relati vely high wages, internaI
structures, and better working conditions must be attributed
to large Japanese f irms according to the segmented labour
market the ory • On the other hand, small firms must illustrate
low wages, po or employment conditions, and a lack of internaI
administrative methods. If Japan portrays the preceding
traits in large and small firms, one must then conclude that
1
7
Japan's extent of labour market segmentation depends on the
~ize of an enterprise.
In 'ilmost aIl the literature on the Japanese labour
market, it is implicitly assumed that duality exists. My
thesis will attempt to explicitly state Japan's form of
duality. Furthermore, l believe that by examining the
validity of labour market ~egmentation in Japan, according to
firm size, the theory's strengths and weaknesses wil: be re-
evaluated. The theory's flexibility to other situations will
also be questioned. Does the dual theoretical framework only
apply to North America? If the theory can be relevant in the
Japanese context, it will be an indication of the theory's
possible universality in industrialized economies. At
present, the literature on the existence of a dual labour
marKet deals mostly with the United States. By studying the
Japanese labour market, on a mainly descriptive and
theoretical level, the segmented labour market theory will be
analyzed in a fairly new contexte
8
, 1
CHAPT ER ONE
LABOU~ MARKET SEGMENTATION TREORY
In order to observe the applicability of the dual labour
market theory (or labour market segmentation) on the Japanese
labour market, a thorough examination of the theory's basic
foundations, assertions, and weaknesses need to be presented.
A comprehensive analysis of the segrnented labour market theory
is essential in understanding the variety of differences that
can exist between firm sizes. An examination of the theory's
basic foundations, assertions, and wf!aknesses need to be
presented, even though the latter two chaptr~rs will focus
mostly on the variarlc charact'~ristics between different firm
sizes in Japan. Since Japan is now a world leader in
technology and certain economic indices, it will be us~ful to
observe if the Japanese labour market structure has similar
characteristics to that found in the United states and thus,
following a duality in nature. The exten i: to which the
Japanese labour mar!",et is "dual", wi th an emphasis on f irm
size differentLltion, will be determined by understanding the
theory's drguments and implications.
J
1
9
origins of Labour Market segmentation
The institutionalist view of labour markets was an
alternative approach in understanding the behaviour of labour
markets, during the 1940's, when an increase in unionization
and the use (. t collective bargaining was bej ng observed
(McConnell & Brue, 1986: 406). Clark Kerr was one of rnany
social scientists ta extend the examination of institutional
labour economics, in the 1950'~, by indicating that there now
existed a cl~ar distinction between unionized and non
unionized sectors (Gordon, 19~2: 7). Doeringer and pÎ(lre
(1971) scated that different submarkets were established
comprising different rules and regulations within each labour
market. According ta Doeringer and Piore, the beginning of
labour market segmentation was observed by differences in
technology, skill of warkers, and degree of unionism within
the labour market. consequently, a dual character within the
labour market was inevitable (Gordon, 1982: 7).
Segmented labour market theory acquired popu lar i ty in the
late 1960's and early 1970's and was undertaken as a research
program by many prominent mainstream labour economists.
However, there were two negative assessments about the the ory
(Cain, 1976; Wachter, 1974) that demonstrated certain major
weaknesses with its foundation. cain (1976), for instance,
argued that the segmented labour market hypothesis was
atheoretical and that the ~~pirical methods used tended to
deviate fram conventional methods of econornic analysis
1
10
(Dickens, 1988: 129). Certain critiques of the theory will be
further presented after an examination of the theory's main
assertions.
Even after initial negative responses to the segmented
labour market theory in the mid-1970'S, the theory reemerged
and has been pursued by economists and other social scientists
using modern tools of imperfect information theory and complex
econometrics models (Dickens & Lang, 1988: 129). According te
Dickens and Lang (1988), the dual labour market model
describes the empirical wage distribution more precisely than
does the sin9le labour market model. In sum, the dual labour
market approach can outperform an equally complex single
labour market model and it provides a better explanation for
the wage distribution.
that the labour market
Dickens, however, does not propose
consists mainly of two different
sectors, but that dualism is a useful simplification (Dickens,
1988: 129-134).
Reasons for Labour Market segmentation Theory
Cain (1976) identifies much of the motivation for the
dual labour market ~odel as originating from social policy
issues of the 1960' s. One of the leading reasons was the
des ire to comprehend the persistence of poverty and of racial
and gender inequalities, while the United states was
undergoing expansion in its edur.ational system and training
programs. Moreover, there was concern about the new wave of
1
11
job insatisfaction and the deteriorating work ethic among
different segments of the labour market. According to Cain
(1976), the rise of monopolies and the use of education and
training certificates as a screening device, prompted the
development of a dual approach. Simildrly, Michael J. Piore
(1975) argued that the dual labour market hypothesis was
introduceà in the 1960'5 in order to analyze the labour force
problems of disadvantaged workers in urban settings,
p~rticularly black workers.
One of the original contributors of the dual labour
market the ory is the analysis by Averitt (1968). His analysis
comprised of a duality in the economy as a whole. Averitt
(1968: 7) stated that the distinction between the primary or
"center" economy versus the secondary or "periphery" economy
can be described as follows. The center economy consists of
enterprises that are large in size and influence. The
organizations within the primary sector are corporate and
bureaucratie. Its production processes are vertically
integrated through ownership and control of critical raw
material suppliers and product distributors. Firms operating
in the center econorny are diversified into many industries,
regions, and nationR. These firms serve national and
international markets using advanced technoloqical equipment
of production and distr ib1ltion.
1
12
The periphery sector is comprised of relatively small
firms which are usually owned by an individual or a family.
The firm's sales are accrued in restricted local markets.
Production technology and uses of marketing are quite
primitive (Averitt, 1968: 7).
Anderson (1987: 573) points out that in the primary
sector, turnover is slow because workers become locked into
specifie career patterns or interndl labour markets.
Advanc~ment is possible only through specifie channels and
promotion is more a function of seniority than productivity.
Anderson classifies secondary sector occupations as unskilled,
requil'ing very little firm-related training. These jobs offer
no career ladders and cohesive bargaining groups do not form
within the company. Furthermore, Anderson indicates that neo-
classical theory is applicable to the secondary segment
because wages are determined competitively.
IDstitutionalist Th.ory
Orthod0X (neo-classical) economics has always emphasized
a competitive wage setting mechanism. In such a situation,
the supply of labour along with the demand for labour
determine a market wage. The institutionalist theory has an
alternative way in understanding the determination of wages.
Collective wage determination under unions is fundamentally
different from competitive wage setting, and this difference
,1
13
is one of the major distinctions be~ween the two approaches.
According to institutionalists, the d~velopment of collective
bargaining has resulted in an administered wage rate rather
than a market rate. Wages are determined by the conscious
decisions of labour and management representatives and not by
impersonal competitive market forces of supply and demande
Institutionalist the ory states that large enterprises and
their unions attempt to alter market forces and usually have
the power to do so (McConnell & Brue, 1986: 406-407).
Internal Labour Markets
In order to understand the dual labour market model, an
examination of the "internaI labour market" concept is
essential. One can argue that the dual labour market theory
is an extension of internaI labour market hypotheses, Once
again, neo-classical theory focuses on a market in which
workers are openly competing for a job and firms are competing
to attract workers. Orthodox theory assumes that the firm
poses no barrier to the competitive pressures of the labour
market. consequently, it is assumed that each wage rate for
each type of worker, within the firm, syrnbolizes the
determination of market forces (McConnell & Brue, 1986: 409).
Institutionalists disagree with the precedlng assertions.
They point to the existence of internaI labour markets as an
example of non-market forces. An internal labour market is an
administrative unit, within which the pricing and allocation
1
,
14
of labour is g~verned by an array of administrative rules and
procedures rather than by economic variables (Doeringer &
Piore, 1971: 1-2). Many enterprises have a hierarchical
system. Each job hierarchy consists of many different
positions within the firm, which is known as a job ladder.
Usually a new worker has had to compete for a "port-of-entry"
position which is found at the bot tom of the job ladder. In
the preceding situation, the neo-classical argument of market
forces and perfect competition may be applicable. Who qets
hired and at what wage rate may very weIl be determined by
market mechanisms for port-of-entry positions. The firm is
competing with other firms for port-of-entry workers since
the y are hirinq the same kind of labour. Institutionalists,
however, argue that for other jobs within the jOb ladder, it
is administrative rules and procedures that determine the wage
and not market forces. Thus, within an internaI labour
market, it is institutionalized rules and procedures, along
with custom and tradition, which are the lnfluential factors
in determining how workers are allocated in the job hierarchy
and what wage rates they are paid (McConnell & Brue, 1986:
410) •
Why do internaI labour markets exist? Firms usually face
enormous costs in the recruitinq and training of personnel
and, therefore, these costs can be reduced by lowering labour
turnover. certain jobs require specifie skills and the use of
specifie technology. Workers aequire the specifie training
15
while working on-the-job. The cost of such training 1s borne
by the employer. In order for the employer to obtain a return
on this investment in human capital, the employer must retain
these workers of specifie trainin', over a period of time.
The implementation of a job ladder is the core characteristic
of internaI labour markets and an attempt to achieve the
desired workforce stability (McConnell & Brue, 1986: 411-412).
Neoinstitutionalists
As Anderson (1987) indicates, there have been three
dominant views on the determination of wages and the cause of
wage inequality in the United states since the Second World
War. F irst, the "neoinsti tut ional ists" emphasized barr iers to
mobility of workers across sectors due to the growth of the
contemporary enterprise, government regulation, and unions
(Anderson, 1987: 571). Neoinstitutionalists assume that
labour is homogeneous but immobile and therefore, the
characteristics of the industry determine wage rates.
Consequently, wage inequality persists because workers
participate in different industries. Accordillg to
neoinsti tutional ists, the most important character istics which
determine wage rates are the degree of unionization, the
profit rate, the concentration of the industry, and capital
intensity (Anderson, 1987: 571).
16
Human capital Theory
The second dominant view has been the human capital
the ory . Human capital advocates focus on worker heterogeneity,
including different investments in human capital, as the
significant cause of wage differentials. Within the labour
market, one worker may earn more than another because he is
more highly skilled. Each worker is endowed with certain
skills because of his genetic make-up or the early childhood
investment by his parents (Anderson, 1987: 572). Skills can
also be accumulated through voluntary training. According to
human capital theory assertions, ln the long-run, impediments
to mobility are se en as relatively insignificant in
determining wage rates. The third influential theory,
according to Anderson, i5 the dual labour market approach
which extends the concepts developed by the preceding two
v iews. However, before embarking on an analysis of dual
labour market theory (or labour market segmentation), a
historical perspective of the origins of a dual approach will
aid in understanding the development of labour market
segmentation.
~istorical Process of the American Labour Market
Gordon, Edwards, and Reich (1982) state that an adequate
dual labour market theory must take into account the
historical process of the American labour market (Gordon et
al., 1982: 2). Gordon's and his associates' analysis was an
"
17
attempt to understand the continuous political and economic
divisions that exist among workers in the united states. They
argue that segmentation in the labour market is due primarily
to historically created divisions. They emphasize the
importance of examining the historical evolut:on of the labour
movement in order to understand the segmp.ntation process that
occurred in the 1950's and still exists today (Gordon, 1982:
1-17) .
First, Gordon et al. present information on the period of
initial proletarianization from the 1820's to the 1890's.
They describe this period as one in which workers had to se11
their labour power to an employer for a wage. Wage labour
became the predominant way of organizing production. Gordon
states that during this period, control over the work process
varied considerably among employers. Sorne employers directly
supervised the organization of work while others left it to
the workers' discretion. Gordon indicates, however, that
between 1820 and 1890, competition played an insignificant
role in determining wages.
The second phase c\')mprise r,:4 t~e homogenization per iod,
from the 1870's to the Second World War. During this period,
many jobs were transformed to a common, semi-skilled type of
function. Moreover, employers and their foremen had most of
the control over the labour process. There was direct
supervision and skills were mu ch less controlled by workers.
During the homogenization period, the labour market became
1 18
more generalized which involved more competition (Gordon,
1982: 3, 14-15).
The segmentation period, according to Gordon et al., from
the 1920' s to the present, brought forth three distinct labour
markets. certain significant characteristics that accompanied
the segmentation period were the rise of structured rules and
incentives and general skills were now increasingly obtained
through education. Relations between employer and employees,
in many cases, were now a reflection of collective bargaining
agreements. According to Gordon, the segmentation of labour
created divisions among American workers which inhibited the
growth of a unified working-c~ass movement (Gordon, 1982: 15-
16) •
Dual Labour Market Theory
After a brief historical development, an analysis of dual
labour market the ory assertions will now be presented. In
arder to understand the dual labour market the ory 's
applicability to the Japanese labour market, the theory's main
theses must be examined. The following is an appraisal of the
dual labour market theory.
Labour economists have extended the internaI labour
market concept and envision two distinct types of labour
markets: the primary and the secondary labour markets. This
notion is labelled the theory of labol.r market segmentation or
the dua 1 labour market theory. Primary labour markets are
19
characterized by large corporations that provide employment
stability and job security, high and rislng wage rates, and
the presence of job ladders with clearly defined opportunities
for occupational advancement. These corporations use
relatively advanced, capital-intensive technology.
Furthermore, there is usually a strong and effective labour
union that represents the workers. The secondary labour
market is identified by small or medium-sized competitive
fLems that are predominantly labour-intensive and usually
require only general training. The secondary sector is
character ized by employment instabi li ty and high turnover, low
wage rates that are relatively stagnant, few opportuni ties for
advancement, production technology that 15 quite out-of-date,
and unions are absent (Piore, 1975; McConnell & Brue, 1986;
Gordon, 1982).
Dickens and Lang (1988: 129) indicate that the primary
sector which is descr ibed by having good working conal tions
and stable empleyment, also is assumed to offer signif icant
returns te human capital variables such as education and
experience. The low-wage secondary sector is expected to
consist of the opposite returns to such variables. Moreover,
primar~ jobs are rationed. Not all workers who are qualified
for primary sector jobs and des ire one, can obta in one.
Therefore, the dual labour market model (seqmented labour
market model) is also a depiction of the differences in the
incarne distribution, the absence of market clearing, and a
20
radical alternative from the basic neo-classical assumption of
fully rational actors and exogenously determined preferences
(Dickens, 1988: 129).
upper-tier. Lower-tier vs. Secondary
Many social scientists would agree that dividing the
labour market into only two parts is not purely accurate.
Economists and sociologists have identified two distinct
components for the primary labour market. According to most
labour market segmentation theorists, the primary labour
market consists of a lower and upper··tier in order to
distinguish the different types of labour found in this sector
(Anderson, 1987; Phillips & Phillips, 1983; Gordon, Edwards,
& Reich, 1982; Piore, 1975).
piore feels tilat the three unique labour segments are
associated with the sociological description of the middle
class, working class and Iower class subcultures.
The upper tier, of the primary sector, mainly contains
professionai and managerial occupations. These jobs are
characterized by high and privately negotiated wages, economic
stability and status, and greater opportunities for promotion
(Anderson, 1987: 573; Piore, 1975: 126). The upper tier,
similarly to the secondary sect or , has high turnover patterns;
however, turnover ill the upper-tier is associated with
advancement, whereas in the se~ondary sector it is associated
with employment instability. Upper-tier jobs, aiso like those
1 21
of the secondary seetor, are distinguished by the absence of
the eIabo~ate work rules and formaI administrative procedures
which are found within the lower-tier segment (Piore, 1975:
126). FormaI educdtion seems to be a pre-requisite for those
wishing to enter the upper-tier of the primary labour market.
Upper-tier work, according ta Piore, seems to offer much
greater variety in the tasks involved and individual
creativity and initiative is encouraged. Moreover, in the
upper-tier, training is to be deduced from a set of general
principles and mobility chains are created in order for an
employee to develop and apply these pl.' inciples.
In the lower-tier, which also includes reiatively
"better" jobs than those found in the secondary sector, on
the-job training is important. The Iower-tier consists rnainly
of production workers. Training tends to be more specifie in
the Iower tier than in the upper-tier. The learnirlg process
is designed to mold a worker into a specif ic set of
surroundings in whïch the same behavioural patterns arise
continuously (Anderson, 1987; Piore, 1975). Thus, jobs in the
lower-tier can become repetitious and routine in their n,~ture.
Wages, as in the upper-tier, are high and usually negotiated
over groups of workers in sorne forro of collective bargaining.
Furthermore, formaI ed~cation is not as important for lower
tier jobs and educational requirements are often overlooked.
Phillips and Phillips (1983) argue that it is primarily
white males who are hired into upper and Iower-tier primary
,1
22
occupations. The secondary sector seems to comprise primarily
young workers, women, or people from minority groups. Such a
finding emphasizes the concept of segregation between the
sexes within the labour market which will not be a main focus
of this analysis of labour market segmentation and its
applicability to Japan.
Labour Market segmentation
Before embarking on a discussion of certain outcomes and
criticisms of the dual labour market approach, a presentation
of Hodson's and Kaufman's analysis will provide the
predominant theoretical assertions of labour market
segmentation. According to Hodson and Kaufman (1983), there
are four basic elements in the dual model. First, they
identify a dual economy which represents the organizational
structure of capital. This consists of the distinction of the
core and periphery economies in the private sector (Hodson &
Kaufman, 1982: 728).
Secondly, the dual economy results in the existence of a
dual labour market which represents the organization of labour
within capital structures. Thus, corresponding to the core
and the periphery sectors respectively are the primary and
secondary labour markets. According to O'Connor (1973), the
most critical distinction between the two labour markets is
the extent to which employment is relatively stable and secure
in the core sector. This distinction based on stability has
23
often been conceptualized in terms of internaI versus external
labour markets (Doeringer & Piore, 1971).
The origins of dual labour markets being associated wi~h
two economic sectors can be understood by observing the profit
maximization strategies of the two labour processes.
According to Hodson and Kaufman (1982: 730), theL - are two
different yet related arguments about how the drive toward
profit maximization produces differentiated labour markets.
The first argument focuses on technology and skill
specificity. In the core sector, job skilis are specifie to
the firm (or the industry) and jobs are hierarchically
arranged in order for current jOb skills to be a prerequisite
for later job skills. This occurs becë:'use of the advanced
technology being used by core firms. since skill specificity
is a requirement, on-the-job training is usually the method
used for workers to acquire these job skills. On-the-job
training increases costs if labour turnover is frequent.
Thus, an enterprise in the core is induced to limit labour
turnover, thus limiting the high costs required for training
individuals into specifie jobs. Consequent ly, in order to
limit labour turnover, a core firm must provide job security
through high wages, fringe benefits, and better working
conditions. Periphery firms, on the other hand, employa more
out-dated method of production and therefore require skills
that are usually at a general level. Training is limited and
inexpensi ve. Consequently, companies in the per iphery do not
....
,
1
24
have to implement pol icies that induce worker stabili ty
(Hodson & Kaufman, 1982: 729-730).
The second argument focuses on the issues of labour
control and cJ ass consc iousness. This argument examines
reasons other than short-run cost eff iciency for the existence
of differentiated labour markets. Edwards (1979) indicates
that there have been three forms of labour control that have
been historically used by business. Simple or direct control
relying on personal supervision has characterized the
periphery. Technical control built into the organization of
work tasks and bureaucratie control built into the
organization of work rt:les have been associated with the core
(Hodson & Kaufman, 1982: 730). The emergence of different
control structures occurred as a response to differences in
organizational size, in the utilization of technology and in
resistance from workers via unions. The technical and
bureaucratie control structures are mainly ways of
coordinating work and delegating author i ty. Hodson and
Kaufman (1982), however, also argue that these control
structures promote fragmentation of the working class within
the f irm as weIl as between f irms by establ ishing artif icial
gradations among different parts of the working class. Such
fragmentation contributes to lon~-run profit maximization by
undermining the development of working class organizationo
that attempt to raise the relative wages of their workers.
The third element of the dual labour market approach,
25
accarding ta Hodsan and Kaufman, consists of three sets af
outcames for workers. First, the extent or likelihood af
mobility between sectors. Second, the conditions of labour in
terms of var ious tangible and intangible rewards. Third, the
existence af workers' cansciousness.
As Thurow (1975) and Doeringez- and piore (1971) have
argued, entry into the primary sectar of the labour market is
governed by a queue. A warker' s place in the queue is
determined hy the cost of training them as permanent
emplayees. Consequently, many workars find themselves in a
vicious circle if they begin their employment in the secondary
sector. The secondary labour market imposes unstable work
histories on workers. Therefore, the preliminary screening,
eliminates such candidates. These undesirable workers have no
opportunity of gaining entrance into core f irms. Gordon
(1972) argues that the persistence of poverty among certain
groups can be explained by the blockage of mobility between
the primary and secondary labour markets. Workers in the
primary labour market, for instance, have a degree of security
which shields them from the possibili ty of downward mobility.
Even when core ernployers face an economic downturn, they
usually layaff employees rather than terminate their
employment in order to protect the firm's investment in the
employees' training (O'Connor, 1973).
There are aiso significant differences between the twa
sectors in terms of the general conditions of labour. Jobs in
26
the primary sector have high rewards such as wages and fringe
benefits being kept above the ('"r:npetitive rate in order to
ensure low turnover. continuing sorne form of education is
encouraged and usually financially supported by core firms.
The opposite is true for companies within the secondary sector
(Hodson & Kaufman, 1982: 731).
The third outcome of the dual
consciousness is fragmentation which
structure
inhibits
on worker
collective
organization. Between the two sectors, there is competition
for the better opportunities and working conditions of the
pr imary sector workers. Such an environment creates different
political associations between the two types of workers.
Moreover, there is further fragmentation in a core f irm due to
the superficial hierarchy which is imposed by the bureaucratie
system of control (Edwards, 1979).
The final element of the dual labour market the ory ,
according to Hodson and Kaufman, is concerned with the social
division of labour- Several motivating factors for the dual
approach were: the persistent poverty of certain social
groups, racial wage inequalities, and gender wage differences
(Gordon, 1972). The basic argument i s that the observed
rac ial and gender differences are caused by the structure of
labour markets. For instance, blacks and other minorities are
very likely to begin their employment in the periphery sector.
Their empl 'Jyment in the periphery gives them an unstable work
history which core employers use as evidence that they are not
..
27
good workers. Therefore, they are not given access into the
core enterprises (Hodson & Kaufman, 1982: 731).
Gordon (1972) presents an argument that explains why
minorities are usually found within the periphery. According
ta Gordon, the jobs geographically avallable to ghetto
minorities are almost always wi thin the periphery. In
addition, the dual labour market model borrows from the neo
classical economic paradigm the idea of direct employer
discrimination as an explanation of minorities being sorted
out into the periphery. According to Becker (1971), many
employer~ have a "taste" for discrimination but only employers
within the pr~mary can afford Sl.lch discrimination due ta their
excess profits. It has also been suggested that core employers
may use race as a convenient screen ta separate workers who
are inexpensive ta train from those who are expensive ta
train. From the preceding asserti ons, radical theor ists have
concluded that the capital ist class fosters disunity and
fragmentation for the working class (Bonacich, 1976).
Finally 1 an examlnation of certain views of labour market
dualism and accompanying criticisms will provide a more
complete depiction of the theory .
28
Evolution of Dualism
Why has labour market dualism evolved? One explanation
is that dualism in labour markets is a reflection of dualism
in product markets (McConnell & Brue, 1986: 419). Su ch a
reasoning can be placed within a historical centext where
labour market dualism is closely related to the evolution of
the business sector of our economy. As Galbraith envisioned p
the business sect or di v ided into the planning (or primary)
sect or and the market (or secondary) sector. The firms in the
primary sector are protected by certain barriers to entry and
have attained considerable monopoly power. They are also
capital-intensive and technologically sophisticated and enjoy
economies of scale with high rates of profit. They are
insulated from competition by other firms and are thus,
concerned with planning for long-run security and growth.
These enterprises attempt to overcome the uncertainties of the
market. They shape and stabilize consumer demand through
ddvertising and other ferms of product promotion. Foreign
markets are controlled through direct investment in foreign
productive facilities, participatjon in cartels, and in
favourable public policies (McConnell & Brue, 1986: 420).
As we have observed, in the secondary sector firms face
strong competition, economies of scale are few, technology is
relatively primitive, and are sUbject to the uncertainties of
market forces. Anderson (1987) states that there is little
empirical testing of the existence of two sectors as
29
hypothesized by dual labour market advocates. Anderson ci tes
Osterman':; analysis which used the 1967 survey of Economic
Opportunity to classify jobs into prirnary and secondary
sectors. Osterman sUbjectively assigned jobs to sectors based
on factors judged to be important by dual labour market
theory. To observe whether the determination of waqes
differed between sectors, he estimated earnings equations for
each sector and concluded that in the secondary sector,
earnings depended only on the amount of time worked. In the
upper tier of the primary sector, education had a positive
effect on earnings while in the lower tier, institutional
constraints were important (Anderson, 1987: 573). Anderson's
study presented the following conclusions. First, the wage
results are consistent not only with the dual labour market
model but also with neo-classical assertions. Wages were
found to be highest among professional workers, managers,
sales workers, foremen, craftsmen, and workers in the
protection field. Most of the high wage jobs are white collar
with the exception of forernen and craftsmcn (Anderson, 1987:
579) •
support for Duality
The fOllowing are the results on the existence of job
traits compared to the predictions of the dual labour market
theory. Anderson's study found that the "better" jobs are in
the primary se-::tor (upper and lower tiers) and are
30
characterized by high wages, specialized training, and
infrequent unemployment. These occupations include white
collar workers (professionals, sales workers, clerks) and blue
collar workers su ch as foremen. Second, the "bad" jobs are
found in the secondary sector and consist of low wages, no
training, frequent unemployment, and instabi lit Y .
displaying most of these trait~ are usually
Occupations
blue collar
( labourers, opera tors, etc. ) .
regression results provide sorne
theories (Anderson, 1987: 584).
criticisms of Duality
consequently, Anderson' s
support for segmentation
According to Hodson and Kaufman (1982), the most damaging
critique to the dual theoretical approach is that a systematic
model has never been presented. Therefore, the research in
this field remains ambiguous and vague since there is no
concrete theoretical framework that one can use. Another
criticism is that the dual model analys is is more descriptive
than theoretical (Hodson & Kaufman, 1982: 732).
According to Piore (1983: 249) , labour market
segmentation theory has met much criticism because it deviates
from the field of "normdl" economies. As piore indicates, the
notion of labour market segmentation was mainly introduced
through participant observation. However, the field of
econornic theory has always attempted ta create a theory which
is continuous and homogeneous in its description of economic
behaviour.
J1
Thus, it is not surprising that labour market
segmentation, which has similarities with marxist economics,
is met wi th opposition. At the heart of labour market
segmentation lies the theoretical notion that the labour
market is not continuous and is heterogeneous. P iore
concludes that labour market segmentation does not fit with
the conventional paradigm of economic theory (Piore, 1983:
250-252) . At the core of labour market segmentation are
social groups and institutions. The processes governing
allocation and pricing within internaI labour markets are
social, opposed to competitive forces. The marginal labour
force cornrnitment of the groups which creates the possibility
for a functional secondary sector of a dual labour market is
social. The hierarchies that distinguish professional and
managerial workers from other employees of the workforce are
also social. Consequently, labour market segmentation requires
a social as opposed to an indi vidual, theoretical econornic
frarnework (plore, 1983: 252). For the purposes of analyz ing
and understanding the Japanese labour market, the descriptlve
distinction of different firrn sizes in "dual" labour market
theory will be utilized as a reference point.
32
CHAP'r!!R TWO
THE THREE PILLARS AND WAGES
Japan - Economie Suceess
After the Meiji Restoration o~ 1868, Japan began to take
interest in the industrialization process. Very few countries
have had such a rapid industrialization as Japan experienced
throughout the late nineteenth and into the twentieth century.
Moreover, faw countries have had su ch economic growth. Even
though the Second World War devastated the Japanese p.conomy,
post-war economic growth has been particularly outstanding.
Today, with a population of over 120 million, Japan is the
world's second largest economy in terms of GNP and among the
richest in GNP per capita. From 1963 to 1973, Japan
experien~ed the highest level of growth in GNP, approximately
10 percent per year. Japan has also registered the lowest
unemployment rate, among industrialized nations, between 1970
to 1989. In the 1980's, the highest unemployment figure for
Japan was 3.1 percent in 1987. Japan' s total labour force has
grown from 57 million to 62 million between 1980 to 1989
(World Labour Report, 1989). Remarkably, the unemployment
rate has remained quite consistent from 1.13 percent 01 the
labour force in the early 1980' s to 2.3 percent in 1989
(McMillan, 1984: 3; Mitsuru, 1990: 106). Since the late
33
1970's, Japan has continuously had a low unemployment rate
compared to other industrialized nations. In the U.K., for
example, the unemployment rate in 1984 was 12.4 percent and
9.5 percent in the united states. Japan, therefore, has
become a leader in economic growth in the 1970's and 1980'5.
One can attribute Japan' 5 sllccess to i ts labour management
organization.
The following two chapters will present the Japanese
labour market characteristics in order to evaluate the extent
to which they are applicable to the labour market segmentation
theory. There will be a particular focus on Japan's firm size
differentiation and its effects. After the analysis of each
Japanese labour market variable, it will be shown that the
Japanese labour market does reveal a dual nature and that firm
size plays a signif icant role in determining employrnent
characteristics.
In this chapter, the three pillars, as the y are usually
referred to, will be presented. Lifetime employment,
enterprise unionism, and the seniority-based wage system wiJl
provide significant evidence of Japanese labour market
segmentation with an emphasis on firm size differentiation.
Moreover, a discussion of wages in various industries will
further demonstrate Japan's dual labour market. As Sasaki
(1981: 17) indicates, the rapid growth of the Japanese economy
has resulted in uneven growth creating two sectors. One is
characterized by big businesses and the other consists of
1
1
34
small and medium-sized companies. Several others have also
supported the idea that Japan has a dualistic labour market
structure (Kalleberg, 1988; Hashimoto & Raisian, 1985; Sasaki,
1981). The information presented will focus on the private
sector since the dual labour market theory's assertions mainly
apply to that sector.
The analysis of Japan's dual labour market will begin by
examining what Many authors refer to as the three pillars.
According to Weiler (1986), the three solid foundations of
certain labour-mdnagement interaction are lifetime employment,
enterprise unionism, and the seniority-based wage system.
Koike (1988: 4-7) describes the existence of the three pillars
in a similar fashion but presents them as a reflection of
Japanese cultural traits, since they are usually viewed as
being unique to the Japanese labour market. Furthermore,
Koike argues that empirical research is lacking in order to
assess how unique the se traits are to Japan. Nevertheless, as
Dore (1989: 26) states, these three labour market systems are
often examined as Integral to the Japanese labour market. By
examining each system, a clear observation on the existence of
dual structures will begin to become apparent.
Lifetime Emplo~ment
The first pillar that will be analyzed is the practice of
lifetime employment. As Weiler (1986: 116-117) points out,
this practice applies to the majority of aIl regular workers
in major enterprises.
35
According to Sasaki's findings, in
1975, 70.7 percent of workers were in enterprises with less
than 300 employees and 29.3 percent of workers were in
companies with over 300 employees (Sasaki, 1981: 22). In
1990, it remains that only a minority of workers are employed
in large enterprises comprising of 500 or more employees.
Lifetime employment, therefore, is applicable only to a
minority of workers. Lifetime employment is the inherent
expectation that once one is hired into a large enterprise, he
will work in various positions within the firm until
retirement. It does not apply to employees hired on a
temporary contract nor to workers who j oin a f inn late in
their career. Consequently, women who interrupt their
employment in order to have a family do not enjoy lifetime
employment. The practice of lifetime employment became
wldespread only after WWII due to a shortage of workers.
Firms were willing to offer lifetime employment in arder to
ensure that their skilled workers would remain with them.
Since management, in large enterprises, undertdkes tremendous
costs in providing on-the-job internaI training, the
implementation of lifetime employment attempts to ensure that
workers will remain with the firm (Weiler, 1986: 117). Thus,
lifetime employment commitment provldes a sense of long-term
membership in the firme Managers and employees have a similar
assumption about their employment contract. They feel like
permanent members of the company and understand that
.
36
throughout the course of their employment, they will perform
different tasks and will be asked to do whatever is needed
(Dore. 1989: 51). Lifetime employrrent reinforces an employee
identity that is closely associated with the firme
According to Dore (1989: 12), since lifetime employment
is prevalent only in Japan's major enterprises, it means that
administrat ive processes play a much more signif icant role
than market forces in employment of workers in large firms.
As Dore indicates, lifetime employment is achieved through
internaI redeployment and intergenerational transition.
InternaI redeployment consists of the practice of large firms
to post individuals to new positions within the firm by the
personnel administration. Intergenerational transition refers
ta the practice of large indi vidual f irms that promote the
process of generation transition by early retirement schemes
ta speed the termination of obsolescent skills, while
maintaining a high level of recruitment of younger workers
(Dore, 1989: 11). InternaI redeployment and intergeneration
transition are practices that reflect administrative processes
rat.nel' than ::lxternal market forces. A 1967 survey illustrated
that 90 percent of firrns with more than 1000 workers had a
formaI retirement scheme, which is only required if a firm is
operating under a lifetime employ~ent guarantee system (Dore,
1989: 56). Only 50 percent of companies with 30-99 employees
had such a system. Sy 1985, almost 100 percent of firms with
over 1000 workers had a formaI retirement plan as compared to
~, ,
li
J7
80 percent of companies with 30-99 workers, and 96 percent of
firms with 100-299 employees. Firms that have fewer than JO
employees presumably have no such retirement schemes (Dore,
1989: 56). It is apparent from these results that the larger
the firm, the more diffused is a system of retirement which is
an indication of a lifetime employment practice.
Lifetime employment ie:; the norm for high school and
college graduates hired straight out of school by major
corporations. Mass hiring has its roots in the practice of
lifetime employment (Yogi, 1988: 197). Major Japanese
corporations have in-house training programs. Thus, when they
hire employees, they do not only look at thei~ existing work
skills but at their potential. The firm wants to develop an
employee thë".t will be a supporti ve team player. An employee' s
lifetime earnings at a company can total between 200 and 300
million y~n when the practice of lifetime employment i5 in
place. Consequently, recruiting an employee is a long-term
investment and a company's future depends on hiring the best
possible workers.
When examining the Japanese lifetime employment practice,
it is apparent that it follows the labour market segmentation
ideology. Dualism emphasizes the different characteristics
that exist in f irms dependlng on the size of the f irm.
Lifetime employment schemes are only observed for
approximately JO percent of the workforce that is employed in
companies with over 300 employees (Sasaki, 1981: 22). Small
38
enterprises do not have th~ means to of fer such a program of
employment. Moreover, lifetime employment in large
enterprises is achieved tt.rough a!l internaI redeployment or
promotion system. As Dore (1989: 11l points out, such a
system is a reflection of administrative forces rather than
market forces. The predominance of administrative forces in
large enterprises is another clear assertion of labour market
segmentation theory. Only employees who are wi thin the
internaI labour market of a large firm enjoy lifetime
employment (Weiler, 1986: 119).
Enterprise Union1sm
The second pillar of the Japanese labour market refers to
enterprise unionism. Rather th an being organized along craft
or industrial lines as in North America, 95 percent of
Japanese industrial unions are found along enterprise
boundaries (Weiler, 1986: 115; McMillan, 1984: 157). J~pan's
enterpr ise unions require that members of the union and union
representatives must be employees of that enterprise.
Nevertheless, Japanese unions are not only conf ined to the
enterprise level. Koike (1988: 249) identifies three levels.
First, there is a labour organiz~tion in separate plants and
companies (i. e. enterprise union). The craft occupations have
organized on occupational and regional bases. The plant
enterprise union is a formaI organization similar to the U.S.
locals. Second, enterprise unions have joined together to
1
•
39
form federations which more or less are affiliated with a
particular industry. These industrial federations are similar
to the national unions found in the United states. At the
third level of organization, these industry wide federations
are affiliated with national labour federations. Sohyo
(General Council of Trade Unions of Japan) is the largest and
accounts for 37 percent of aIl union members (Facts and
Figures of Japan, 1985; Koike, 1988). Domei, Churitsu Roren,
and Shinsanbetsu are among the other major national labour
federations. There has recently been a move towards
unification of the nationwide organizations.
In Japan, however, the foundation of labour organization
is the enterprise local or ~ranch union. As McMillan (1984:
157) points out, in small firms with one plant there exists
one local. In large f irms, the locals combine te form a
company-wide body. In large Japanese companies, unjons are
primarily interested in lifetime employment stability. Wage
increases are determined mu ch more as a function of enterprise
productivity and national wage norms which will be discussed
later. The enterprise union, according to McMillan (1984:
157) and Koike (1988: 252), is involved in the management and
decision making within the company. Moreover, management in
Japanese companies often comes from union ranks. Membership
in the union local is available to aIl white-collar and blue
collar employees. Since management positions above the
position of section chief are excluded from the union, most
1
,
40
managers in the hierarchy who have attained positions above
section chief have moved through union ranks and were probably
engaged as an official in the union. Since mobility across
firms is minimal due to lifetime employment in large
companies, enterprise unions help create allegiance of
employees to the union and to the company (Mf;Millan, 1984:
158) .
In 1975, more than 90 percent of aIl unions in Japan were
enterprise unions and 80 percent of aIl union members were
affiliated with enterprise unions. Unionized workers
represented approximately 36 percent of total employed workers
in 1964, 32 percent in 1975, and 29.7 percent in 1983 (Shirai,
1983: 122). The constant decline in union membership is
similar to that experienced in Canada and the United states
throughout the 1970's and 1980's. Japan's labour union
membership of 29.7 percent in 1983 is almost half of that in
the United Kingdom which stood at 57 percent in 1981 (Facts
and Figures of Japan, 1985). It is accurate, therefore, to
conclude that only a minority of employees are members of
enterprise unions in Japan. Furthermore, it is evident that
membership in unions clearly depends on company size (Koike,
1988; Shirai, 1983; Sasaki, 1981). The estimated rate of
union organization in the private sector, for aIl industries,
by size of establishment is as follows. In establishments
with 29 or less employees it is 3.4 percent, for 30-99
employees it is 9.0 percent, for estatlishments with 100-499
1
41
workers it is 31.5 percent and for 500 or more employees it is
approxitnately 63.6 percent (Sasaki, 1981: 50). Thus, the
larger the establishment, the higher the proportion of
membership in an enterprise union with the accompanying above
mentioned characteristics. In Shirai (1983: 103), it is
confirmed that trade-union membership is far less apparent in
smail firms than in large ones. Shirai indicates that in
firms with 49 employees or less, trade-union membership is
approximately 13 percent. In companies with 300 or more
workers, it reaches over 65 percent. Koike (1988: 199) also
states that labour unions are prevalent in large enterprises
but only 27 percent of smal1 and medium-sized companies had
labour unions.
As was found for the lifetime employment scheme,
unionization largely depends on company size. The dual labour
market the ory emphasizes that the primary sector which i5
comprised of large f irms i5 much more unionized than the
secondary sector which contains smaller firms. In Japan, the
prece~ing assertion is exactly what we find. As Sasaki (1981)
and Shira1 (1983) present, mernbership in Japan's enterprise
unions is much higher in firms employing over 500 employees
than in smal1er companies (approximately 64 percent for
companies with over 500 employees and 3 percent for firms with
29 workers or 1ess). consequently, the size of an enterprise
is a significant determinant of employee unionization. The
larger the firm, the higher the percentage of union activity.
1
42
seniority-Based Waqe system
The third pillar in Japan's labour-management framework
is the seniority-based wage system (Nenko). In large Japanese
enterprises, a worker is recruited after secondary or
university education and begins employment at a relatively low
starting salary. The workers low starting salary is related
to their age and education, rather than to ability. Each
worker' s wage increases continuously until retirement (Weiler,
1986: 119). The advantage of the Japanese seniority-based
wage system is that while a worker's income increases, it does
not only provide him with incorne security but also coincides
with his needs as he gets married, has children, and so on.
As weIl, favouritism is elirninated and workers have a
disincentive to leave the firm because they will have to start
at the beginning of the wage system, once hired by another
firm (Weiler, 1986: 119). As stated before however, the
seniority based wage system is found in large enterprises.
only a minority of workers are employed in these firms,
therefore the majority of the workforce does not face a
concrete seniority-based wage hierarchy.
Waqe Data
An analysis of wage data is required in order to evaluate
where the seniori ty-based wage system is operating. By
exarnining wage differentials for blue-collar workers, white
collar workers, males, and females it will become apparent
43
that the seniority-based wage system and higher wages are
instituted in large companies.
Kalleberg's (1988: 12l) analysis provides evidence that
organizational structures in Japan create dlfferences ln
employees' earnings. As Kalleberg points out, there are
certain basic differences in the process of earnings
determination in the United states and Japan. According to
his findings, American manufacturing workers' salaries are
predominantly determined by job characteristics, positions in
the authority hierarchy, and union representation. The
salaries of Japanese workers are mostly shaped by life-cycle
variables such as age and by aspects of an internaI labour
market, for instance, seniority and internaI training.
Blue-Collar Workers
Koike' s (1988) study reinforces the structural seniori ty
based wage system as an important factor in determining a
worker's earnings. Koike indicates through his findings
(1988: 20) that there are significant wage differentials in
the distribution of earnings of fuale blue-collar employees.
The findings reveal that the wages of male blue-collar workers
in companies with 1000 or more employees appear to increase
rapidly along with age. For instance, the wage index which lS
set at 100 for workers in the 21-24 age group rises sharply to
over 160 for those in the 35-39 age group and to approximately
170 for employees between 50-54 years of age. In small firms
J 1 ...
44
with 10-99 employees, wages do increase with age but only
slightly. The peak is when the wage index reaches 140 for
workers between 35-39 years of age. After the 35-39 age
group, there is steady decline. In large firms (1000 or more
employees), there is a strikingly steep drop in the wage index
among the oldest blue-collar workers. The index drops to
approximately 105 for those aged 60 years and over, which is
closely associated to the 21-24 age group. When this is
compared to the peak of 170 for those between 50-54, it is
remarkably low. since retirement at many companies commences
at 55, recently increased to 60, it is understandable that
anyone working over the age of 55 begins receiving lower
wages. In enterprises with 10 to 99 employees, a similar
decline in the wage index is apparent, however, it is less
steep than in large Japanese firms. There are several
assertions one can present according to Koike' s findings.
First, it is clear that companies with 1000 or more employees
pay higher wages to their workers compared to firms with 10-99
employees. The wage index (Koike, 1988: 20) is situated
higher on the vertical axis for companies with 1000 or more
workers. Secondly, it is apparent that a seniority-bdsed wage
system has been implemented in large Japanese firms. The wage
index demonstrates a continuous rise in wages from the 18-20
age group until the 50-54 age group. As mentioned previously,
the only surprising aspect is that wages drop significantly
for workers over 55 years of age in large Japanese companies.
J
45
As Koike asserts, perhaps there is a contra-seniority wage
system that operates for older workers.
White-Collar Workers
Koike (1988: 29) also studied wage differentiais of
white-collar workers against biue-collar empIoyees. Koike's
study compares blue-collar workers in companies with 1000 or
more empIoyees and white-collar workers in companies with 10-
99 employees. The findings conclude that monthly wages for
Japan's white-collar workers in firms employing 10-99 workers,
are significantly higher than those for biue-collar workers in
large firms of over 1000 employees.
that in the above situation it
Such a finding indicates
is not the size of the
enterprise but rather the type of worker that determines an
employee's position in the wage scale. Nevertheless, as Koike
(1988: 20) pointed out, Japan's blue-collar workers in firms
with more than 1000 employees have higher wages than blue
collar employees in medium-sized or small enterprises.
Moreover, there seems to exist a seniority-based wage system
for white-collar males as has been indicated for biue-collar
males. For white-collar males in companics with 10-99
employees, there is a continuous rise in wages, with a peak of
approximately 210 in the wage index for those around 50 years
old, except for the decline for those over the age of 54. The
preceding pattern is almost identical ta that of biue-collar
males cited previously in Koike (1988).
46
According to Koike (1988: 53-54) and Kalleberg (1988:
127) through their findings, there is evidence that wage
differentials between blue-collar and white-collar workers are
more narrow in Japan than in any ether industrialized nation.
However, as Koike indicates, the narrow wage differentials
between these two groups is largely found only in large
Japanese companies. Koike' s results seem to demonstrate that
being a wh i. te-collar or a blue-collar worker does not indicate
higher wages. On the contrary, it is working for a large firm
versus a srnall firm that determines wage rates.
Male Earnings
Mitsuru's (1990: 108-109) finding also clearly port ra y
that pay raises by seniority is a significant trait of the
wage payment system in Japan. For instance, average monthly
contractual earnings for males in aIl industries indicate a
constant increase ~"1S age increases from the age one enters the
labour force to the 50-54 age group. Wages for the 50-54 age
group are over 450 thousand yen per month in sharp contra st to
workers in the 20-24 age group who received around 180
thousand yen per month. Moreover, Mitsuru's findings present
a significant wage disparity between mal~s working in
companies with 1000 or more employees and those working in
firms of 10-99 workers.
'l'able 2. 1 :
Montbly Male Earninqs· by Firm Size for 20-24 Age Group
Number of Employees
10-99
10U-999
1000 and over
Monthly Earning~
170 000 yen
175 000 yen
200 000 yen
Monthly Male Earninqs· by Firm size for 45-49 Age Group
Number of Employees
10-99
100-999
1000 and over
• Monthly Contractual Earnings
Monthly Earninqs
300 000 yen
360 000 yen
450 000 yen
Source: statistical Handbook of Japan, 1990, p. 108-109.
47
The above f iqures which represent the labour market
situation in 1988 are very similar to those found by Koike
(1988) for the years of 1976 to 1979. There is no great
disparity between sizes of f irms for age groups under the age
of 24 which indicates that for entry positions workers face
similar wages regardless of firm size. For age groups over
the age of 24 however, the wage differentials between
different firm sizes are consistently significant as those
discussed for the 45-49 age group. For employees in the 45-49
age group, there is a wide wage differentiatial between firm
1
sizes.
48
There are several conclusions that one can assert
after examining these f indings. First, even though workers in
the 20 to 24 age group receive similar starting wages
regardless of firm size, it is apparent that the larger the
f irm, the higher the monthly wages paid in each age group.
Secondly, as age increases, the wage differentials between
different firm sizes becomes more significant. According to
Mitsuru (1990), larger firms pay much higher wages than
medium-sized firms at each point along the age-wage profile.
similarly, in medium-sized f irms employees receive higher
earnings than workers in small firms. consequently, the size
of the firm seems to significantly determine the wages workers
will edrn. Thirdly, the highest differences in wages between
age groups is recorded within firms employing over 1000
workers (Mitsuru 1990: 109). For instance, in firms with over
1000 workers those in the 50-54 age group receive 460 thousand
yen per month while those in the 20-24 age group earn 200
thousand yen; in f irrns wi th 100-999 workers those in the 45-49
age group get 360 thousand yen per month as compared to 175
thousand yen in the 20-24 age group; in firms with 10-99
employees thosp. in the 45-49 age group earn 300 thousand yen
per month in contra st to workers in the 20-24 age group who
receive 170 thousand yen (Mitsuru, 1990: 109). It is evident
therefore, that the seniority-based wage system is far more
implemented in larger firms than in smaller ones.
49
Female Earnings
Alrnost sirnilar conclusions, as the ones stated above, are
found in Mitsuru's (1990) findings on female monthly
contractual earning<ç, by scale of employrnent. As is the case
for males, the larger the firm the higher the earnings fOl"
females.
Table 2.2:
Famal. Monthly contractual Earnings
Company size
10-99
100-999
1000 & Over
Age Group
25-65
50-54
60-64
Peak in Salary
150 000 yen
180 000 yen
325 000 yen
Source: statistical Handbook ot Japan, 1990, p. 109.
The table clearly indicates that the larger the firm, the
higher the monthly earnings for females. simi lar to the
findings for males, Mitsuru's (1990: 109) charts indicate that
larger firms pay fernales rnuch higher wages than smaller firms
at each point along the age-wage profile. Furtherrnore, the
greatest wage differentials between age groups within the same
type of firm was found in cornpanies employing over 10UO female
workers which was also found for males. In these f irrns,
ernployees begin their rnonthly earnings at around 120 thousand
-
50
yen in the 15-19 age group and reach 325 thousand yen in the
60-64 age group. Between these age groups, a constant
earnings increa~e is observed. In these firms, the large wage
disparity between age groups indicate the existence of a
seniority-based internaI wage system in large companies. In
medium-sized enterprises employing between 100-999 female
workers, monthly earnings begin at around 110 thousand yen and
reach 180 thousand yen for those in the 25-29 age group. From
the 25-29 age group until the 55-59 age group, monthly
earnings remain fairly constant around 180 thousand yen.
Wages slightly decline after 59 years of age. Therefore,
medium-sized enterprises do not demonstrate any seniority
based compensation system since monthly earnings remain at
approximately the same level between 25 and 59 years of age.
The same is true for small enterprises employing between 10-99
female workers. Monthly earnings begin at around 135 thousand
yen for the 20-24 age group and rernain virtually constant at
150 thousand yen per month for aIl the remaining age groups.
It is clear that in both medium-sized and small enterprises,
females do not face a seniority-based system. Their wages do
not increase with age. Men, on the other hand, do face wage
hikes as their age increases regardless of firm size. The
largest disparity between age groups for males, however, is
found within cornpanies that employ over 1000 workers.
Consequently, for males and females, a seniority-based wage
system seerns to be L.lplemented predominantly in large f irms
51
validi ty of the labour market segmentation theory. rrhe the ory
clearly identifies that large firms contain internaI seniority
based wage systems. In Japan, it is clear that the larger the
firm, the more widespread is a seniority-based wage system.
Koike (1988) clearly indicates that when considering blue
collar workers, the larger the firm the more likely that an
age-wage relationshlp exists. He found the same was true for
white-collar workers. Mitsuru (1990) asserts that for males
and females monthly earnings also seem to follow a seniority
based wage system in large firms. For females, it is only in
firms with over 1000 workers that wages rise as age increases.
For males, in aIl flrms there is an increase in wages as age
increases but the largest dispar i ty between age groups is
found in firms employing over 1000 workers. In both cases, it
is apparent that a seniority-based wage system is predominant
in large firms. These findings totally support the labour
market segmentation theory which distinguishes sma Il f irms
that do not have an intricate seniority-based wage system from
large firms that do possess elaborate wage programs based on
seniority. Moreover, Kalleberg (1988: 125) supports the
notion that earnings differentials by firm size are greater in
Japan than in any other industrialized nation.
The labour market segmentation theory aiso asserts that
the larger the firm, the higher are the wages paid to the
employees. The Japanese labour market depicts this preceeding
1 theoretical assertion.
52
Sasaki (1981: 21) points out that
wages in the manufacturing industry solidify the notion that
the larger the firm, the higher the wages. Firms employing
between 50-99 workers paid 62.3 percent of what firms with
over 1000 workers paid in wages. Firms with 100-499 workers
paid 74 percent and with 500-999 workers paid 85.8 percent in
comparison to companies with over 1000 employees. Mitsuru's
(1990) and Koike's (1988) wage data illustrates that larger
firms pay higher wages to males, females, blue-collar, and
white-collar workers. Almost identical results were recorded
by the Facts and Figures of Japan in 1985. According to
Sasaki' stables, approximately 70 percent of the workers
working for small and medium sized firms get 60 percent or
less of employees' wages in large firms. As Sasaki indicates,
this is evidence of a dual structure in the Japanese economy.
A few large companies with high wages coexist with numerous
small and medium-siz~d firms that pay low wages (Kalleberg,
1988: 145; Sasaki, 1981: 20) .
• age setting
One of the most exceptional aspects of the Japanese
industrial relations system is the manner in which wage
increases are handled. Wage increases are determined at the
industry level for 72 percent of the workforce (Dore, 1989:
48). The practice of annual wage hike negotiations began in
Sohyo in 1956 and extended to other labour centres su ch as
1
53
Domei. By 1970, this practice of wage setting became known as
the spring labour offensive or Shunto (Weiler, 1986: 124).
During Shunta, the industry-wide suggested wage increase that
is arrived at by federations, is used as the standard for
negotiations that take place at the enterprise level. The
final decision on wage increases rests with labour management
at each company (Weiler, 1986: 124). payment of a large part
of the total wage, however, is in the form of a bonus. In
manufacturing firms with over 500 workers, 30 percent of the
total wage is given in bonuses; 18 percent for f irms wi th 10-
99 employees (Dore, 1989: 42-43). Larger firms pay d higher
proportion of wages in the form of a bonus perhaps because of
an internaI performance evaluation system. Dore (1989: 48)
suggests that it is large firms that influence the outcome of
Shunta and that it is usually empIoyees of large firms that
are directly affected by the spring labour offensive. Weiler
(1986: 136) also asserts t.hat not aIl employees' wages are
determined by Shunta. According to Weiler, many workers who
are employed in the sccondary labour market must re ly on
market forces to determine their wage raises. Whereas, for
large firms, Shunto is an indication of an internaI
administrative process which is insul ated from smaii
competi tive enterpr ises.
InternaI administrative forces demonstrate credibility
for labour market segmentation arguments. Once again, large
firms follaw an institutionalist approach to wage
,
54
determination. As the segmented labour market the ory asserts,
an internaI process of wage increases is a trait of large
firms in the primary labour market in comparison ta the small
f irms in the secondary labour market which are dictated by
market forces.
55
CHAPT ER THREE
L~RGE VERSUS SMALL FIRMS
Women in the Japanese Labour Market
As in most industrialized nations, Japanese women are
beginning to enjoy certain improvements in the realm of female
employment. Women's share of wage employment in 1980 was 34.1
percent and increased slightly to 36.5 percent in 1987 (World
Labour Report, 1989). Education has been a signif icant
impetus for advances in women's labour market activity. A
growing proportion of wornen, for instance, cont inue from lower
secondary school to upper secondary school. Over 50 percent of
females continue to upper secondary education in comparison to
80 percent of males. In 1976, only 13 percent of the fema le
population attended university (41 percent of males) and
approximately 20 percent were enrolled in Junior Colleges
(only 2.4 percent of males). There are more than twenty
million women in the Japanese labour force which accounts for
46 percent of the total female population (McMillan, 1984:
135). According to McMillan's findings, 70 percent of women
in Japan are employed in small or medi um-sized companies.
They are predominantly found in productjon and clerical jobs
where mobility i5 limited, employment benefits are minimal,
and retirement is about ten years earlier than men.
1
1
56
consequently, even though there have been advances in women' s
labour force participation and educational attainment, women
are in a relatively inferior position, as compared ta men, in
almost aIl areas of employment. As Dore (1989: 32) indicates,
women are used as shock absorbers responding ta the business
cycle. In 1985, women accounted for 4 percent of regular
part-time workers and 3.6 percent of temporary part-time
employees (as compared to 2.2 and 0.9 percent respecti vely for
men) . On] y 23.9 percent of women were employed in regular
full-time positions as compared to 56.6 percent of men (Dore,
1989: 33). It is apparent that Japanese women play a
significant role in part-time work.
The annual wages or salaries of regular employees in
companies with 30 workers or more averaged 5.2 million yen for
men and 2.6 million yen for women in 1989 (Mitsuru, 1990:
108) . Women are receiving only 50 percent of men's wages,
which is the widest wage differential by se~ of any
industrialized economy. Koike (1988: 46-47) also illustrates
that in manufacturing, when the male blue-collar wage is set
at 100, blue-collar wages of females in their thirties and
forties lie below half of the corresponding male wage. White
collar female wages, in the 30-34 age group, lie at
approximately 53 percent of the corresponding male wage.
Above 34 years of age and until retirement, white-collar
females receive less than half of the male wage. The same
trend ~s found for blue-collar females. White-collar females
57
receive around 85 percent of the rndle wage in the 21-24 age
group and then exper ience a constant decl ine. Blue-collar
fernales receive approxirnately 69 percent of the male wage in
the 21-24 age group and rnuch lower wages for aIl the older age
groups. It is apparent, therefore, that for middle-aged and
older fernale workers the wage differential by sex is extremely
wide (Koike, 1988: 46). This pattern indicates that a
seniority-based wage system is not irnplemented for wornen. Dore
(1989: 74) reinforces the notion that a seniority-based wage
scheme is absent for wornen. Furtherrnore, Dore (1989) presents
data on starting wages in Sakai city for newly hired
employees, as seen in the following table.
Table 3.1:
startinq Waqes in Sakai
Age Se~
20-29 Male
35-54 Male
20-29 Fernale
35-54 Fernale
Source: Dore, 1989, p. 74.
MonthJy Salary
173 000
216 000
130 000
116 000
Dore asserts that the sarne pattern is found in sales. With
Koike's (1988) and Dore's (1989) findings, it is clear that
1
1
58
men are introduced into a seniority-based wage system even in
small enterpr ises. Women, on the other hand, in aIl f irm
sizes, are paid less at age 35-54 than in their 20's.
Koike (1988: 44-45) presents wage differentials for women
by company size in Japan.
Table 3.2:
White-Collar Female Waqes by Firm size
Number of Employees
10-99
100-999
Wage Index (%)
89
91
Blue-Collar Female Waqes by Firm Size
Number of Employees
10-99
100-999
Wage Index (%)
75
82
Wage Index for Firms wi th over 1000 employees = 100.
Source: Koike, 19H8, pp. 44-45.
As Table 3.2 indicates, different sizes of firms produce
varying wage scales. It is evident that a larger firm pays
higher wages. In examining the situation of women in Japan,
segregation and discrimination between the sexes is obvious.
Women, in 1989, were only receiving 50 percent of men's wages
(Mitsuru, 1990: 100). Wornen, at every age group are
compensated much less than men.
59
\'Jage differentials between
male and female manufacturing workers ouer the age of 30 are
exceptionally wide (Koike, 1988: 46). Women receive less as
t.hey get older, unlike the seniority-based wage scheme that is
found for men (Dore, 1989: 74). The preceding results
illustrate that women's equity i5 far from the Equal
Employment opportunity Law enacted in 1986 by the Japanese
government (Yogi, 1988: 198).
In observing how labour market segmentation applies to
the situation of women in Japan, one must argue that certain
theoretical assertions are not found. Women employed in large
Japanene firms do not enjoy the benefits of the core sector.
The seniority-based wage system, for instance, seems ta only
apply for men and not aIl workers in the primary sector.
Koike' s {1988} resul ts however, demonstrate that labour market
segmentation's notion that different labour market traits are
found in different firm sizes does gain credibility. Wornen's
wages are higher the larger the firm (Koike, 1988: 45).
Moreover, as labour market segmentation implies, Japanese
women are seen to be primarily employed in small firms
(McMillan, 1984).
Education
Educatllm in Japan seems ta play an important role in
determining the type of employment opportunities one can
attain. Many Japanese companies recruit recent university,
60
college, and high school graduates once a year to begin work
usually on Apr il 1 (Yogi, 1988: 196). starting salaries
remain quite similar across industries. The best graduates
who enter enterprises wi th over 1000 employees start on
average at 147 000 yen per month. New gradua tes going to
f irms that employ 10-99 workers start at approximately 140 000
yen monthly (Dore, 1989: 67). Even though starting salaries
are quite uniform regardless of firm size, larger companies do
pay higher st3rting salaries than smaller firms. Moreover,
Japanese white-caIlar college graduates in companies with over
1000 workers receive higher wages, in the dge group of ~arly
thirties until retirement, than white-collar high school
gradua tes in f irms wi th over 1000 employees (Koike, 1988: 32).
Thus, a higher level of education does provide higher wages,
while keeping other variables constant. Koike (1988) also
presents the percentage distribution of male executives by
rank, by age, and by educational grade in large companies with
over 1000 employees. In the age group 30-34 years of age,
university gradua tes accounted for 35 percent of executive
positions whereas junior college graduates and high school
leavers accounted foc approximately 20 percent of executive
positions. In the 35-39 age group, almost 75 percent of'all
occupations at the executive level were held by university
graduates, 40 percent by college graduates, and 37 percent by
high school leavers. In the above two age groups, the
positions held at the executive level were subsection heads,
61
deputy heads of section, and section heads. No one had
achieved the position of department head which is the highest
position at the executive level.
In the 40-44 age group, approximately 12 percent of
department heads were university educated as compared to )
percent being occupied by college and high school graduates.
When aIl executive positions are considered, university
graduates hold over 80 percent, college graduates
approximately 55 percent, and high school graduates around 50
percent (Koike, 1988: 216). These results indicate that a
seniority-based promotion system, according to age and
education, has been implemented in large Japanese companies.
As Koike points out, there are no department heads in the age
groups of 30-34 or 35-39. There are a few department he~ds in
the ~0-44 age group but most are in the age group of over 45
years. Furthermore, educational qualifications do play a
significant role in determining the attainment of executive
occupations. It is evident that a higher proportion of
uni versi ty and college graduates hold executi ve posi tians than
high school leavers (Kaike, 1988: 216). Hashirnoto and Raisian
(1985: 726-727) also illustrate that where education is
concerned, 25.7 percent of ernployees in large t irrns had
completed a coilege education as compared to 22.2 percent for
medium-sized firms and 11.9 percent for small firms.
Labour market segmentation theory asserts that large
enterprises that provide gaod employment benef i ts have a
1 62
higher proportion of college or university graduates as
employees. The f indings above reinforce these theoretical
statements. Furthermore, top executive positions are usuaIIy
held by college or university educated workers, which is
directly applicable to the description of the upper-tier of
the primary labour market of labour nlarket segmentation.
Kojke's findings on the attainment of executive occupations
also emphaslzes the existence of an internaI seniority-based
promotion system which also applies to the core sector of
labour market segmentation theory. Koike (1988) implies that
such a system is not found within smaii cornpanies. Hashimoto
and Ralsian's (1985) results illustrate that the proportion of
employees with a college or university education in large
f irms is double that found in small enterprises. This
indicates that the completion of post-secondary education is
more of a necessity in entering a large firrn than a small one.
Once again, it is observed that different sizes of firrns
require different employee credentials in order to construct
different promotion ladders.
Large vs. Small Firms
It is already apparent that there are Many different
labour market characteristics that sharply distinguish large
firms from smaller ones. It is in the large firrns that the
distinctive features of the Japanese labour market seem to
exist. Lifetime employment, enterprise unions, and the
63
seniority-based wage system are far more widespredd in firms
with over 1000 employees than in companies with 10-99 workers.
Moreover, it is predominantly male school graduates, that are
recruited bo: large firms that use a very selective screening
process, who work in jobs that are seniority-based in their
benefits (Kalleberg, 1988: 125). According to Kalleberg, it
is large firms that invest heavily in internal training,
welfare services, and social plugrams for their employees.
Kalleberg (1988) and Koike (1988) suggest that small firms pay
lower wages, provide fev/er welfare services, and limited
opportunities for promotion. Furthermore, small firms are not
likely to practice nenko. Small Japanese companies are often
subcontractors of large firms.
Kalleberg (1988) also demonstrates that the internaI
labour market system is mostly found in very large
enterprises. The nenko wage system is one of the pr imary
characteristics that is embedded in the internaI labour market
concept. The internaI labour markets of large f irms are
further characterized by careful screening of potential
workers, high jOb security, frequent job rotation, extensive
internaI training, internaI career mObility, and a seniority
based compensation and promotion system.
Even though there are major differences in employment
characteristics depending on firm size, there is an
interdependence bet ween the two. Cha l mer s (19 8 9) f ee l s t:, ~ L
many studies on labour markets ignore the relevance of the
1
,
64
peripheral sect or which is mainly comprised of small firms.
Chalmers suggests that the high wages and good working
conditions of large firms depend on the low wages and marginal
employment benef i ts of sma 11 companies. Thus, an
interdependence between different firm sizes exists. From the
preceding discussion on large and small firms, it is evident
that labour market segmentation assertions are being followed
in the Japanese labour market. Gordon, Edwards, and Reich
(1982) and Hodson and Kaufman (1982) emphasize the different
employment characteristics between the primary and secondary
labour market enterprises. In order to clearly demonstrate
that Japan's labour market follows different employment traits
depending on the size of a firm, an appraisal of certain
labour market characteristics will follow.
Internal Promotion
Koike's (1988: 116-117) findings indicate that the larger
the company, the higher the proportion of occupations that
benefjt from internaI promotion. The internal promotion
system is extensively found in the ~hemical industries where
it encompasses 80 percent of occupations in firms with over
1000 employees. The same is found in the primary metals
industries. 1:, the machinery and light industries, the
internaI promotion system is not as widespread. According ta
Koike (1988), the internaI promotion system prevails in large
Japanese companies far more than it does in the united states,
across aIl industries.
65
An executive position in large
Japanese companies is largely unattainable without having a
given number of years of continuous service to the company.
There are very few department or section heads with less than
10 years of service. Only 8 percent of section heads have
serviced between 5-9 years. These figures rise tremendously
once an employee surpasses 10 years of continuaI service.
Over 50 percent of section heads are in the group of 10-19
years of service. The next attainable position is to become
a department head, which usually requires over 20 years of
continuaI service (Koike, 1988: 211). These results give
support to the existence of a seniority-based promotion of
large Japanese companies. Koike (1988) impl ies that such a
system is not found ira small firrns. In Japan, the internaI
promotion system dppears to be widespread thlC~ghout aIl the
industries and a high correlation is found between large
Japanese companies and internaI promotion schemes. According
to Koike (1988: 280), the internaI seniority-based promotion
structure is far more prevalent in large Japanese companies
than in large firms in Western European nations or the United
states. The internal promotion system, which is widespread in
large rather than small Japanese companies, strengthens the
labc~r market segmentation affirmations that diffe:ent
characteristics accompany different sizes of f irms. Labour
market segmentation cIearIy identifies internaI mechanisms of
1
1
66
promotion in the primary labour market which primarily
consists of large enterprises.
InternaI Training and Mobility
Koike (1988) and Shirai (1983) suggest that the larger
the firm, the more prevalent is an internaI training program.
In large companies with over 1000 employees, workers acquire
a wider range of skills by rotating jobs within one workshop.
After several years of service, they move to other workshops
with the same technological sophistication. Thus, an
employee's experience and skill is enhanced in large Japanese
enterprises. Small Japanese firms do not have the proper
resources to offer such elaborate internaI training programs.
Shirai (1983: 97-98) also specifies that the larger the firm,
the greater is the use of internaI mobility. Approximately,
75 percent of companies with over 300 workers do provide the
internaI training and rnobility mentioned above. Only 35 to 45
percent of smaller f irms offer such an internaI program.
Furthermore, surveys on the use of outside courses were mostly
used by small f irms. Off-the-job courses were used by 50
percent of firms with 30 to 99 workers but only 24 percent of
firms with over 1000 employees (Dore, 1989: 53). These
results reflect the capability of large firms to provide
internaI rather than external training. Dore (1989: 40-41)
also points out that sorne large firms provide a dual career
track. They offer a "comprehensive career" track for sorne
employees and a "general duti"s" path for others. This device
67
is used primarily to attract graduates from top universities
in the former career track while hiring less able graduates,
or women, into the latter. In one sample survey, 32 percent
of firms with over 5000 employees had such a system, but only
3 percent of aIl firms. Once again, the Japanese labour
market lends support to the labour market segmentation
ideology. InternaI training is predominantly found operating
in large firms. Small firms depend more on external training •
mechanisms.
Job Tenure and Separation Rates
When aIl non-agricultural private industr ies are
considered, job tenure is 13.9 years for firms with over 1000
workers, 10.2 years for firms with 100-999 employees, and 8.5
years for small companies with less than 100 workers
(Hashimoto & Raisian, 1985: 726-727) . These results
demonstrate the tendency for longer job tenure to occur in
large firms. Workers stay longer with larger firms presumably
because the employment benefits are better.
In manufacturing companies with 1000 or more empioyees,
the annual separatioll rate for males in thei . early twenties
is more than 20 percent. For males in their thirties and
forties, in large companies, i t is only 6 percent. It is
apparent that the permanent ernployment scheme is in force for
workers in their thirties and forties, in firms with over 1000
employees (Koike, 1988: 57) . Dore (1989: 56) also exhibits •
1
T ,
68
the diffusion of the lifetime employment system, in large
Japanese firms, by observing separation rates.
Table 3.3:
Average Annual Separation Rate by Firm size (in percentaqe)
Number of Employees
5-29 28 21
30-99 36 19
500 and over 24 14
Source: Dore, 1989, p. 56.
Consequently, separation rates depend on company size.
Smaller firms have higher separation rates. According to Dore
(1989), the difference in separation rates between large and
small firms is explained by the greater probability of small
f irm ernployees to qui t rather than by a tendency for employers
of srnall firms to disrniss their workers. It is evident that
better ernployment benefits in larger firrns, as described by
dual labour market allegations, causes separation rates to be
lower and job tenure to be longer.
69
Workweek and Overtime
The length of the workweek depends significantly on the
size of the firme Approximately 54 percent of workers, in
firms with less than 100 employees, worked six days every
week. Whereas 25.5 percent of workers in firms with 100-999
employees and only 3.5 percent of workers in companies with
more than 1000 workers were required to work six days a week.
The rnajority of employees in medium-sized and large firms work
a standard of five da ys per week. For small companies, the
regular work week consists of six days (Facts and Figures of
Japan: 1985).
The requirement to work overtirne, on the other hand, is
primarily observed in large firms. Employees work 22 hours
overtime per week in cornpanies with more than 500 workers, in
comparison to 15 hours in firms with 30 to 99 employees (Dore,
1989: 29). According to Dore (1989), the lifetime employment
expectation in large firms makes employers more l ikely to
resort to overtime than to hire more employees. When
examining the workweek and overtime habits of Japanese
workers, i t is clear that laL-ge f irms have :l dit ferent
structure of employment than small companies. Since a
difference depending on company size exists, labour market
segmentation assertions are once again upheld.
70
~ality Control circles and Joint Consultation
Quality control circles are informaI discussion meetings
that are conducted outside working hours, in which workers
make ;uggestions in order to improve their job, work area, and
the production process. Usually, a quality control circle
consists of eight to ten employees and one supervisor who
works in the same area (Koike, 1988; Weiler, 1986; McMillan,
1984). Quality control circ les are implemented by management
in order to induce workers to make suggestions on possible
ways of improving productivity. Such a work group involves
the active participation of aIl employees and promotes
eff icient labour-management relations, worker self-development
and teamwork (Koike, 1988: 152; McMillan, 1984: 158). Koike
(1988: 8-9) infers that quality control circles are
predominantly found in large enterprises where there is a
dependence on the workers' technical skills.
As part of the collective bargaining process, there are
joint consultation sessions between union representatives and
management, usually once a month. These meetings provide the
opportunity to discuss issues relating ta the introduction of
new technology, productivity goals, safety, fringe benefits,
grievances, and so on (Koike, 1988: 252; Weiler, 1986: 129).
In most medium-s i zed and large f irms, workers on the shop
floor are given the privilege of discussing a production plan,
being proposed by upper-Ievel management, during the joint
consultation sessions (Weiler, 1986: 130). Thus, joint
71
consultation is a means of exchanging information and
attempting to reach an agreement on how the firm should be
rnanaged. According to Koike's (1988: 252) findings, joint
consultation meetings are primarily found in companies that
have labour unions. Joint consultation is found in
approximately 90 percent of firms with over 300 employees and
in 70 percent of companies with 100 to 299 workers that also
have a union. Even though the joint consultation system is
predominantly implemented in enterprises with a labour union,
there are several firrns that utilize joint consultation in the
absence of a union. In companies with over 1000 employees and
no labour union, joint consultation is found in 50 percent of
the cases. For smaller companies, it approaches )) percent.
Quality control circles and joint consultation sessions
are prevalent ln large Japanese companies as opposed to
smaller ones. Moreover, quality control circles and joint
consultation reflect internaI structures rather than market
forces. Labour market segmentation advocates pronounce the
distinction of internaI operating mechanisms in large
enterprises and the use of market forces in small companies.
The findings above clearly exhibit that it is larger companies
that tend to implement administrative structures.
72
Productivity DifferentiaIs
since large companies use various internaI methods in
organizing their workforce, it is clear that they have also
succeeded in aChieving productive efficiency. Sasaki's (1981:
20) f indings indicate that large companies are more
productive. A company, with 1-9 ernployees, achieves only 34
percent of the productive efficiency of firms wjth more than
300 employees. Cornpanies w i th 10-99 workers achieve 52
percent, firrns with 20-99 workers reach 56 percent, and
enterprises with 100-299 employees achieve 71 percent of
productivity output that firms with over 300 workers attain.
These results re-affirrn labour market segmen~ation views that
find larger cornpanies, which are considered to be in the core
sector, to be more productiv8 than smaller firrns.
T ,
ft
73
CONCLUSION
As Kalleberg (1988: 125) indicates, Japan's economy has
developed unevenly and is extremely dualistic. For lnstance,
in the manufacturing industry, there are certain large firms
with the latest technological sophistication that coexist with
small companies that use primitive methods. Furthermore,
Kalleberg argues that the emphasis of the dual economy in the
United states focuses predominantly on differences among
product markets and industrIes. Whereas economic dualism in
Japan occurs primarily within industries, among organizations
of different sizes. My analysis' main attempt has been to
demonstrate that labour market dualism in Japan i5 indeed
about the differences of large versus small firms. AlI of the
data gathered has totally reinforced the notion that Japan's
form of dualisJ1l mainly depends on different character istics of
different company sizes. McMillan (1984: 181) also specifies
that a dual sector econorny has persisted in Japan for several
decades. Chalmers (1989: 17) points out that the most
important aspect of industrial dualism lies in the dichotomy
of large and small firms. Once again, the entire analysis of
the Japanese labour market has mainly focused on this
dichotomy. The division consists of a minority of workers
employed in large firms with relatively high wages and good
working conditions and a majority of workers in small
cornpanies with low wages and irregular working conditions. As
74
Sasaki (1931: 22) confirrns, only 29.3 percent of workers are
ernployed in firms with over 300 workers. The rnajority of
employees in Japan work in small enterprises without the high
wages and internaI benefits of a large finn. Lifetime
ernployrnent schernes, seniority-based wage systems and
enterprise unionisrn lS a reflection of only a minority of the
Japanese worktorce. The Japanese labour market, consequently,
dernonstrates significant dlfferences in organization and
emp.l.ùyrnent character istics depending on the size of a cornpany.
The tindings in Chapter Two seern to re-affirrn that the
size of a f irrn plays a sign if icant role in deterrnining the
traits of employment a worker faces. Wages for blue-collar
and white-collar workers are higher in cornpanies with over
1000 ernployees than in srnaller firms. Similar results were
found for rnale and female earnings, across aIl industries.
Lifetirne ernployrnent pra~tices, enterprise unionism, and
seniority-based wdges were predominantly found in large
Jùpanese cornpanies. Furthermore, wage differentials between
age groups were mu ch wider in companies with more than 1000
workers, indicating a solid internaI seniority-based wage
sysLern. The Shunto out corne is also predominantly influenced
by large firms. These preceding findings reinforce the
allegations of labour market segmentation advocates. The
segmented labour market theory clearly identifies high wages
and internaI ernployment structures with large firrns in the
core labour sector. These internaI mechanisms and high
75
compensation packages a exactly what exists in the Japanese
labour market. Lifetime employment, enterprises unions, and
seniority-based wage systems clearly demonstrate internaI
administrative forces in large f irms rather than market. forces
which operate in small companies. The institutional ist
argument of a clear distinction between f irm s izes is once
again solidified.
In Chapter Three 1 the resul ts aga in demonstra ted that
labour market segmentation assertions were prevalent in the
Japanese employment system. Women in Japan, receive only 50
percent of what males earn 1 ind icat Ing discr i mind t ion and
segregation )f women into low wage employment. Most Japancse
women find employment within the sccondary labour mtlrket which
consists of small firms with low wages. This pattern of
female employment i5 exactly what labour market segm8ntation
theorists describe. According to Dore (1989) and Koike
(1988), the primary labour market characteristics don't seem
to apply to wornen even when they are employed by large firms
within the core sector. The senior i ty-based wage system, for
instance, do es not seem to be implemented for women in large
firms. It is primarily men, as labour market segmentation
advocates confirm, that enjoy the full benefits of large firms
in the primary sector.
Where education is concerned, it is evident by the
results, that larger firms hire more of the educated employees
than smaller companies. Moreover, the internaI promotion
76
1 system of large firms allocates top management positions
according to age and educational attainment. Labour market
segmentation arguments infer that large enterprises in the
core sector have a higher proportion of college or university
educated employees thùn small firms in the periphery. Japan's
labour market data re-affirms the preceding argument.
Different employment characteristics depending on company
size were also observed in aspects of internaI promotion,
training, and mobility. Koike (1983) explicitly presented
data that illustrates how widespread internaI promotion
systems are in large f irms. In small f irms, they are virtually
non-existent. Srirai (1983) demonstrated that the larger the
firm, the more likely that an internaI training program
exists. InternaI training in large firms is accompanied by
extensive internaI mobility for employees. InternaI training
and mobili ty, according to the data in Chapter Three, is
rarely found w i thi '1 small companies. Labour market
segmentation theory affirms that internaI organizf.ltion for
promotion, training, and mobility is excIusiveIy established
in large firms of the primary labour market. In Japan, it is
large f irms that predominantly possess internaI operating
mechanisms. Smail companies in Japan seem to rely primarily
on externaJ training. Furthermore, for workers in small
firms, elaborate promotion and mobilit.y opportunities are
limited. Thus, an internaI labour market notion, as is
described by labour market segmentation followers, is dominant
77
in larger Japanese firms.
since working conditions seem to be better in larger
firms, it is not surprising that job tenure is longer in large
Japanese firms and separation rates higher in small companies
(Dore, 1989; Hashimoto & Raisian, 1985). The longer work""eek
in small firms and the higher expectation of overtime
employment in larye f irms portray the di f ferences j n the
structure of labour in different f irm sizes. Moreover,
quality control circles and joint consultation channels are
prevalent in large J apanese compan] es (Ko i ke, 1988). As a
re5ult, productivity i5 much higLer in large firms (Sasaki,
1981). These observable differences between large and small
firms in Japan reinforce the argument of firm size
differentiation in labour market segmentation. Qua l i ty
control circles and joint consultation sessions, for instance,
are indicative of internaI administrative structur8S that are
more likely to exist within larger firms in Japan.
In exanlning the Japanese labour market, it 1S clear that
it has fo] lowed trends in Western economies. The service
industry a:counted for approximately 50 percent of the share
of total f!mployment by 1987 (World Labour Report, 1989: 8).
Dore (1989: 21) indicates that the growth in the service
sector has accompanied an increase in self-employrnent. People
who ar€ self-employed do not face the employment
character istics of employees in large or sma 11 f irms. In
Japan, 13 percent of those in commercial occupations and 30
78
percent of those in service jobs are self -employed (Dore,
1989: 22). Thus, approximately one third of the workforce, in
commercial and service jobs, is not working in large or smaii
companies. The type of employment traits that self-empIoyed
workers face must differ greatly from the internaI benefits,
such as lifetime employment and seniority-based wages, of
large firms. Since self-employment does not adhere to the
primary or secondary labour market organization, a
trichotornous description of the labour market may be
necessary. Further res,earch is required on the employment
characteristics of the self-employed in order to shed more
understanding on the dynamics found in labour market
organization. Research into self-employment may illustrate a
third component of labour market analysis which does not apply
to large or small firm ideology.
My analysis, however, has focused on the differences
between large and small firms in Japan as described by labour
market segmentation the or ists. Japan clearly demonstrates
such a dichotomy within its labour market. There are a
minority of Japanese employees that work in large firms that
offer high wages, lifetime employment, and many internaI
benefits. On the other hand, there are a majority of Japanese
workers who are employed in small companies that offer l.OW
wages and few employrnent bel"efits. As Taira (1986: 35)
indicates, labour market segmentation in Japan largely refers
to two sets of firms, large and smail. It does not primarily
79
refer to two types of occupations, primary and secondary, as
in the United states. Tai ra (1986: ) 5) c lear 1 y states that
the size of a firm in Japan is an extremely powerful element
which determines the behaviour of firrns and employees. My
research has illustrated the significance of company size in
the Japanese labour market. The 1947 Doyukai Declaration
recognized the firm as a coalition of three equals: managers,
workers, and shareholders. It overlookcd, however, the
inequity between workers who are employed in firms of
different size.
1
r \
•
80
BIBLIOGRAPHY
Anderson, Kathryn H., John S. Butler, and Frank A. Sloan. "Labour Market Segmentation: A Cluster Analysis of Job Groupings and Barriers to Entry." Southern Economic JOllrnal, v. 53, January 1987, pp. 571-90.
Averitt, Robert T. The Dual Economy: The Dynamics of American Industry structure. New York: W.W. Norton, 1968.
Becker, Gary, EditiQn.
s. The Economies of Discrimination, 2nd Ch icago: Uni vers i ty of Chicago Press, 1971.
Bonacich, Edna. "Advanced Capitalism and Black/White Race Relations in the United states: A Split Labour Market Interpretation. Il Arneri...:an Sociological Review, v. 41, 1976, pp. 34-51.
Ca in, Glenn. "The Challenge of Segrnented Labour Market l'heor ies to Orthodox Theory: A Survey. " Journal of Econom ic Li tera ture, v. 14, December 1976, pp. 1215-1257.
Chalmers, Norrna J. Industrial Relations in Japan: The Peripheral Workforce. New York: University of Oxford, 1989.
Dickens, William T. and Kevin Lang. "The Re-emergence of Segmented Labour Market Theory." The American Economie Review, v. 78, May 1988, pp. 129-34.
Doeringer, Peter B. and Michael J. Piore. Intern'3.l Labour Markets and Manpower Analysis. Lexington, Mass.: D.C. Heath and Company, 1971.
Dore, Rona Id P. Flex.ibility.
Japan at Work: Markets, Management, Paris: OECD, 1989.
and
Dore, Ronald. "Rigidities in the Labour Markets." Government and Opposition, v. 23, Autumn 1988, pp. 393-412.
Edwards, Richard C. Contested Terrain. New York: Basic Books, 1979.
Facts and Fiqures of Japan. 1985.
Japan: Foreign Press Center,
1
------------------------------------,~~------------........ 81
Galbraith, John K. The New Industrial State. BOston: Houghton Mifflin, 1967.
Gordon, David M. Theories of poverty and Underel!lIÙ.9yment. Lexington, Mass.: D.C. Heath and Company, 1972.
Gordon, David M., Richard Edwards, and Michae l Re ich. Seqmented Work, Divided Workers: 'l'he Historical Tr.:lnsformation of Labour in the l.J.n i t~çL Stat~~. New York: Cambridge University Press, 1982.
Hashimo ta, Masanor i and John Ra is ian. "Employment Tenure and Ea r-nings Prof i le in J apan and the Un i ted States." The Am,erican Economie Review, v. 75, Septembcr 1985, pp. 721-35.
Hodson, Randy and Robert L. Ka ufman. "Economie Oua 1 i sm: A Critical Review." American SociolQ9..ic_~;LRevie_~, v. 47, December 1982, pp. 727-39.
Kalleberg, Arne L. and James R. Lincoln. "The structure of Earnings Inequa li ty in the Uni ted states and J apan. Il American Journal of SociolQgy, v. 94 supp., 1988, pp. S121-S153.
Koike, Kazuo. Understanding Industrial Rcldti<?I)~ ___ in~QQern Ja.Pi!Il. New York: St. Martin's Press Inc., 1988.
Kano, Toyohiro. StrategL and Struct_~r~ __ QÎ ____ Ja.f>i'!nCs_~ Enterprise. Armonk, New York: M. E. Sh<lrpe lne., 1984.
McCOnnE! Il, Campbe Il R. and Stanley L. Brue. Contem129_rarY Lèlbour Economies. New York: McGraw-Hill, 1986.
McMillan, Charles J. The Japanese Indust_rial ~~l~J:!]. New York: Walter de Gruyter & Co., 1984.
Minoru, Morita. "Labour's Realignment Rocks the opposition." Japan Echo, v. 15, no. 2, Summer 1988, pp. 66-70.
Mitsuru, Ide. statistical Handbook Q_f __ Japan. Tokyo: statistics Bureau, Management and Coordination Agency, 1990.
O'Connor, James. rrhe Fiscal Crisis of the State. st. Martin's Press, 1973.
New York:
Phillips, Paul and Erin Phillips. Women and Work: Inüquality in the Labour Market. Toronto: James Lor imer & Company, 1983.
82
Piore, Michael J. "Notes for a 'rheory of Labour Market St rat if i c a t Ion. " l n =L=a:.=b:;..;:o::;..u::.r=---=-M..:.:a=r..:..:k:..::e;..::t'-----=S:;..;:e:::..;g~m=ec:..n:...;:t::.:a::..;t::..;l::.· .:o.:.,:n • Lexinton, Mass.: D.C. Heath and Company, 1975.
Piore, Michael J. "Labour Market Segmentation: To What Parad igm Ooes i t Belong?" The Amer ican Economic Review, v. 73, May 1983, pp. 249-53.
Sasaki, Naoto. Management and Industrial structure in Japan. Oxford: pergamon Press Ltd., 1981.
Shirai, Ta ish iro. Contemporary Industria l Relations in Japan. Madison: The University of Wisconsin Press, 1983.
'raira, KO]I. is it?" 35-37.
"Labour Market Segmentation in Japan: How Rig id Monthly Labour RevIew, v. 109, June 1986, pp.
Thurow, Lester C. Generating Ineguality. New York: Basic Books, 1975.
Wachter, MIchael L. "Primary and Secondary Labour Markets: A Critique of the Dual Approach." Brookings Paper on Economie Activity, v. 3, 1974, pp. 637-80.
WeIler, Joseph M. "The Japanese Labour Relations System: Lessons for Canada." Labour-ManagelJ\ent Cooperation in Canada. Toronto: University of Toronto Press, 1986.
Wohlers, Eckhardt and Gunter Weinert. Employrnent Trends in the United States, Japan, and the European Community. New Brunswick, U.S.A.: Transaction Inc., 1988.
Wor Id Labour Report: 1989. Geneva: International Labour Organization, 1989.
yoj i, Ishikawa. "The 1988 Job Market for Recent Graduates." Japan Quarterly, v. 35, no. 2, June 1988, pp. 196-200.