Sie sind auf Seite 1von 146

Lecture Notes in Economics and Mathematical Systems

(Vol. 1 5: Lecture Notes in Operations Research and Mathematical Economics, Vol. 16-59: Lecture
Notes in Operations Research and Mathematical Systems)
Vol 1 H . Buhlmann, H. Loeffe l, E. Nievergelt, EinfUhrung

In

die

Vol. 30: H. Noltemeier. Sensltlvltatsanalyse bel diskreten Iinearen

The one und Praxie de r Entscheidung bei Unsicherhelt. 2. Auflage.

Optlmierungsproblemen. VI, 102 Seilen, 1970.

IV . 125 Selten. 1969.

Vol. 31 M . Kuhlmeyer. Die nichtzentrale t,Verteliung . II. 1 06


ten. 1970.

Vol 2 U. N. Bhat A Study of the Queueing Systems M/G/1 and


GI/MI1 VIII 78 pages. 1968.
Vol. 3

A Sirauss An Introduction to Optimal Control Theory.

Out of print

Vol. 4. Branch and Bo und : bne F infUhrung. 2 .. geanderte Auf lage.


Herausgegeben von F. Weinberg. VII. 174 Selten. 1973 .
Vol. 5' L P_ Hyvarinen. Informati orl Theory for Systems Eng ineers
VII, 205 pag es. 1968.

Vol. 6 H. P. Kunzl,

Muller, E.

~Jlev e rgelt,

Elnfuhrung skursus ,n

die dynamlsche Programmierung. IV, 103 Selten. 1968.

Vol. 7 W . Popp, Elnfuhrung In die Theone der Lagerhaltung . VI.


173 Seiten. 1968.
Vol. 8 ' J T eghem, J Loris' T eghem . J P. Lambotte. Modele ,
d'Attente MIG/ I et GI/MI1 a Arrlvees et Services en Groupes. III,
53 pages. 1969.

Se l ~

Vol. 32: F. Bartholomes und G. Hotz. Homomorphlsmen und Re '


duklionen linearer Sprachen. XII 143 Seiten. 1970. DM 18.Vol. 33 : K. Hinderer. Foundations of Nonstationary DynamiC fJrogrammlng With Discre te T ime Parameter. VI. 160 pages. 1970,
VoL 34 H. S tormer. Seml -Markoff -Prozesse mit e ndllch vle len
Zustand en Thearie und Anwendungen. VII, 128 Selten. 1970

Vol. 35 : F Ferschl. Markovkett en. VI. 168 Selten. 1970.


Vol36 M . J. P. Magill, On a Generol Economic Theory of Motion
VI, 95 pages. 1970.
Vol 37 H . M U l l er~Merb ac h. On Round,Off Errors In Linear Pro'
gramml ng, V.48 pages, 1970 .
Vol. 38 Statlstl sche Method en I Herausgegeben von E Walter.
VIII. 338 Selten. 1970

Vol. 9: E. Schultze. ElnfOhrung In die mathematischen Grundlagen


der Informatl on stheone. VI, 116 Selten. 1969.

Vol 39 : Statlsllsche Methoden II Herausgegeb en von E. Walter.


IV. 157 Selten. 1970.

Vol. 10: D. Hoc hstad ter Stochastis che Lagerhaltung smodelle. VI.
269 Seilen. 1969.

Vol. 40 ' H . Drygas. The Coordinate'Free Approach to


Markov Estl matlOn. VIII. 113 pages 1970,

Vol. 11 / 12: Mathematical Systems Th eory and Economics. Ed ited


by H. W Kuhn and G. P Szegb. VIII, III. 486 pages. 1969

Vol. 41 ' U. Ueln g, Zwel L6 sungsmethoden fUr nlchtkonvexe Pro'


gramm le rungsprobleme. IV, 92 Selten. 1971

Vol. 13' Heuristische Planungsmethoden. Herausgegeben von

Vol. 42 . A.V. Balakrishnan. Introduction to Optimi zati on Theory In


a Hilbert Space. IV. 153 pages. 1971

F. Weinberg und C

A Zehnder II. 93 Selten. 1969

Vol. 14: Computing Methods


1969.

1'1

Optimizat ion Problems. V, 191 pages.

Gauss~

Vol. 43 J A Morales. Baye Sian Full Information Structural Analy '


SIS. VI. 154 pages . 1971

Vol. 15 : Economic Mode l s. Estimation and Risk Programming

Vol. 44 :' G Felchtlnger. Stoc has tlsc he Made l le demograph ischer


Essays in Honor of Gerhard T,ntner. Edited by K A . Fo x, G V . L Prozesse, IX, 404 Selten. 1971.
Naraslmham and i. K Sengupta . VIII. 461 pages. 1969
Vol. 45 K. Wendler. Hauptau stausc hschntte (Pnnclpal Pivoting),
Vol. 16: H. P. KUnzi und W. Oettli. N,chtl,neare Optlmlerung
II. 64 Selte n. 1971
Neuere Verfahren . Bibliographie. IV. 180 Seilen. 1969.
Vol. 46: C. Boucher, LeGon s sur la theorle de s automates maVol. 17: H. Bauer und K Neumann. Berechnung optlmaler Steuethematlques. VIII , 193 pages. 1971

rungen. Maximumpnn zip und dyna mlsc he Optlm lerung. VIII. 188

Seilen . 1969.
Vol. 18 ' M. W olff . Optlmale Instandhaltungspolltiken
Systemen V. 143 Selten. 1970

In

elnfachen

Vol. 19 L. P. Hyvannen. Mathemati ca l Modeling for Indu stnal Pro'


cesses, VI, 122 pages. 1970.
Vol 20: G . Uebe . Optimale Fahrpl ane. IX, 161 Selten. 1970
Vol. 21: Th. M . Liebling. Graphentheorie In Planungs- und l ourenproblemen am B eispiel d es stadt lsc hen StraBendlen s tes . IX.

118 Seilen. 1970


Vol 22: W. Eichhorn, Theone der homogenen ProduktlOn sfunk,
tion. VIII, 119 Seilen 1970
Vol. 23: A Ghosal. Some Asp ec ts of Queueing and Storage
Systems. IV, 93 pages. 1970

Vo l. 47 H . A Nour Eldin , Optlmlerung l inearer Regelsysteme


mit quadrati sc her Z'elfunktlon. VIII. 163 Seiten . 1971 .
Vol. 48 M. Constam, FORTRAN fUr Anfanger. 2. Auflage. VI,
148 Selten. 1973
Vol. 49 : Ch. SchneewelB, Regelungstechnlsche stochasllsche
Optiml erungsverfahren. XI, 254 Seilen. 1971
Vol. 50 : Unternehmensforschung He ute - Ubers lchtsvortrage der
ZUncher Tagung von SVOR und DGU , September 1970. Heraus'
gegeben von M , Beckmann . IV. 133 Seiten. t971
Vol. 51 Dlgltale Sim ulat ion. Herausgegeben von K. Bauknecht
und W. Net. IV. 207 Selten . 1971
Vol. 52: Invanant Imbedding, Proceedings 1970. Edited by R E.
Bellman and E. 0, Denman, IV, 148 pages. 1971.

V, 66 Seiten , 1970,

Vol. 53 : J RosenmUller, Kooperative Spiele und Markte. III , 152


Selten , 1971

Vol. 25' R Henn und O. OPitz, Konsum, und Produklionstheorie I.


II , 124 Seiten. 1970.

Vol. 54: C. C. vor Welzsa cker, Steady State Capital Theory. III,
102 pages, 1971

Vol. 26: 0 Hochstadter und G. Uebe, Okonometnsche Methoden.


XII. 250 Selten. 1970

Vol. 55: P. A V. B. Swamy, Statistical Inference iQ Random Coef'


ficlent Regression Models, VIII, 209 pages, 1971 ,

Val 27 ' I. H. Muft i, Computotional Methods in Optimal Control


Problems. IV, 45 pages, 1970.

Vol. 56: Mohamed A EI-Hodiri, Constrained Extrema, Introduction


to the Differentiable Case with Economic Applications. III , 130
pages. 1971

Vol. 24 : G. Feichhr'ger, Lernprozes se in stochastischen Automaten.

Vol 28 Theoretical Approaches to Non,Numerical Problem Sol,


vlng. Edited by R B Banerji and M, D. Mesarovic, VI, 466 pages.
t970

Vol. 57: E Freund. Zeitvariable Mehrgr6Bensystem e. VIII,160 Sei'


ten. 1971.

Vol. 29: S. E EI maghraby, Some Network Models in Management


SCience III 176 pages, 1970.

Vol. 58' P. B. Hagelschuer, Theorie der linearen Dekomposition


VII, 191 Seiten. 1971
continuation on page 137

Lecture Notes
in Economics and
Mathematical Systems
Managing Editors: M. Beckmann and H. P. Kunzi

Mathematical Economics

116
K. Miyazawa

Input-Output Analysis and the


Structure of Income Distribution

Springer-Verlag
Berlin Heidelberg New York 1976

Editorial Board
H. Albach' A. V. Balakrishnan' M. Beckmann (Managing Editor)
P. Dhrymes . J. Green . W. Hildenbrand . W. Krelle
H. P. KUnzi (Managing Editor) . K. Ritter' R. Sato . H. Schelbert
P. Schonfeld
Managing Editors
Prof. Dr. M. Beckmann
Brown University
Providence, RI 02912/USA

Prof. Dr. H. P. KUnzi


Universitat ZUrich
8090 ZOrich/Schweiz

Author
Dr.. Kenichi Miyazawa
Hitotsubashi University
Kunitachi,
Tokyo, 1861Japan

Library

or Congress Cataloging

in Publication Data

Miyazawa, Ken' ichi, 1925Input-output analysis and the structure of income


distribution.
(Mathematical economics) (Lecture notes in economics
and mathematical systems ; 116)
Bibliography: p.
Includes index.
1. Interindustry economics. 2. Income distribution-
Mathematical models. 3. Japan--Economic condi tions-Mathematical models. I. Title. II. Series. III. Series: Lecture notes in economics and mathematical
systems ; 116.

HB142.M59

339.2

76-000006

AMS Subject Classifications (1970): 90AlO, 90A15, 90A99


ISBN 978-3-540-07613-1
001 10.1007/978-3-642-48146-8

ISBN 978-3-642-48146-8 (eBook)

This w.ork is subject to copyright. All rights are reserved, whether the whole
or part of the material is concerned, specifically those of translation, reprinting, reuse of illustrations, broadcasting, reproduction by photocopying
machine or similar means, and storage in data banks.

5.

Under
of the German Copyright Law where copies are made for other
than private use, a fee is payable to the publisher, the amount of the fee to
be determined by agreement with the publisher.
by Springer-Verlag Berlin' Heidelberg 1976

FOREWORD

The purpose of this study is in keeping with the shift in concern over the economic problems of growth to those of income distribution in recent years.

Income

distribution problems may be analyzed by not only the traditional procedures, but
also by some extensions of the input-output technique as I shall demonstrate in this
volume of the Lecture Notes.

Some fruitful results are obtained by applying the

extended input-output technique to income analysis as well as to output analysis.


This volume consists of three parts.
with some overlapping unavoidable:

These parts may be viewed along two veins,

(1) Parts One and Two contain extensions of the

input-output analysis and (2) Parts One and Three contain studies of the effects of
the structure of income distribution on some other economic relationships.
First, as an extension of the input-output analysis, we present a synthesis of
the Leontief interindustry matrix multiplier and the Keynesian income multiplier in
disaggregated form, and introduce a new concept which may be called the "Interrelational Income Multiplier" as a matrix.

It is designed to analyze the interrelation-

ships among various income-groups in the process of income formation through the
medium of industrial production activity.

Although this multi-sector multiplier

follows from Leontief's interindustry matrix multiplier, it is formulated by the


inclusion of the income generation process, which is omitted in the usual inputoutput open model, and by projecting the multiplier process into not only the outputdetermination side, but also into the income-determination side.
Secondly, we shall proceed to formulate a method of partitioning off the original Leontief inverse in terms of the combined effects of "Internal" and "External"
matrix multipliers and their induced sub-multipliers.

Because the usual Leontief

inverse provides us with knowledge of only the ultimate total effects of interindustry propagation and not the disjoined effects separable into the partial multipliers,
as such, our method may well be applied to the various kinds of problems that require us to trace back to the interactions among two or more strategic industry
groups.
Finally, some empirical applications of these two models are introduced, deal-

IV

ing with several cases from the Japanese economy and with an international comparison
of the interdependence between service and goods-producing sectors.

The empirical

illustrations also include the applications of an interregional version of the inputoutput model in the extended forms.
The other theme of this volume deals with the structure of income distribution.
In this context, we employ two methods of entirely different nature.
The first is an application of the above mentioned interrelational income multiplier model, by which we clarify the effects of income-distribution-factors on the
income determination process.

In the standard income analysis or in the standard

input-output open model, the same amount of autonomous expenditures cannot have varying effects on the level of national income even if the expenditures consist of different commodity proportions.

The same criticism holds for the Kalecki-Kaldor type

of models -incorporating income-distribution-factors as far as there are no changes in


the relative income shares.

But in the real world this situation is not so.

It will

be shown that in order to have the value of income vary in conjunction with the commodity proportions of demand, it is not sufficient to introduce the structure of
income distribution by types of income-group alone, but we must introduce at the
same time the distribution-factors by the types of industrial value-added for the
production structure also.
The second study is a differentials-analysis especially of wages and interests
as rewards to the factors of production.

In contrast to the above approach which

focuses upon the interindustry intermediate inputs as factors of production, we


concentrate our attention directly on income distribution among the primary inputs
by the size of firms.

The analysis is an integral part of the last Chapter's in-

vestigation of Japan's dualistic structure.

The dualistic character of the Japanese

economy, reflected mainly in production techniques and financial arrangements, are


considered in relation to distribution and to economic growth.
While numerous individuals have made important suggestions and criticisms, I am
especially in debt to K. Ara, A. S. Bhalla. W. H. Branson, S. Masegi, K. Ohkawa.
M. Shinohara. Y. Shionoya, and T. Watanabe.

Special gratitude is extended to Ryuzo

Sato who read through the original manuscript with constructive criticisms and who

v
recommended this volume for publication in this Series.

The author gratefully

acknowledges Gilbert Suzawa for correcting and improving the English content of the
original manuscript.

Lastly, the author wishes to thank the various Journals, as

noted in the footnotes to each chapter, for permission to reproduce the original
articles in various revised form.
Tokyo,
August 1975
Kenichi Miyazawa

INPUT-OUTPUT ANALYSIS AND THE STRUCTURE OF INCOME DISTRIBUTION


BY KENICHI MIYAZAWA
CONTENTS
FOREWORD
PART ONE:

INPUT-OUTPUT AND INCOME FORMATION

CHAPTER 1 INTERINDUSTRY ANALYSIS AND THE STRUCTURE OF INCOME DISTRIBUTION ...... .


I.

Introduction ........................................................ .

II.

Interindustry Analysis and the Process of Distribution and Expenditure


of National Income...................................................

1) The Leontief Multiplier, Keynesian Multiplier and Kalecki Multiplier

III.

2)

Gener&lization of the Input-Output Model

3)

The Coefficients of Inter-Income-Groups

The Relationship of Inter-Income-Groups and the Multi-Sector Income


Multipl ier ...........................................................

1) The Income Multiplier as a Matrix


2) Accepted Multipliers as Special Cases
3)
IV.

Structure of the Propagation Process

The Convergence Conditions of the Model .............................. 14


1) The Properties of Leontief-type Matrices
2)

Convergence Conditions in the Model

CHAPTER 2 INPUT-OUTPUT ANALYSIS AND INTERRELATIONAL INCOME MULTIPLIER


AS A MATRIX ........................................................... 22
I.

Introduction ......................................................... 22

II.

Formulation of a Regional Income Multiplier in the form of a Matrix .. 23

III.

The Interrelational Income Multiplier among Regions .................. 26

IV.

Composition of Final Demand and the Regional Income-Distribution ..... 29

V.

Output Determination and Interregional Income Generation ............. 36

VIII

CHAPTER 3 FOREIGN TRADE MULTIPLIER, INPUT-OUTPUT ANALYSIS AND THE CONSUMPTION


FUNCTION .............................................................. 43
I.
II.

Introduction ......................................................... 43
The Foreign Trade Multiplier and the Circular Flow of Intermediate
Products ............................................................. 44

III.

The Modified Multiplier and the Fundamental Equation for an Open


Economy .............................................................. 46

IV.

Interindustry Analysis and the Consumption Function .................. 48

V.

Empirical Estimates .................................................. 51

VI.

Formula for the Computation of the Subjoined Inverse showing the Effect
of Endogenous Changes in Consumption ................................. 55
1)

Derivation of the Computation Formula

2)

Propagation

PI~cess

Combining Leontief's Multiplier and the

Keynesian Multiplier
PART TWO:

INTERNAL AND EXTERNAL MULTIPLIERS

CHAPTER 4 INTERNAL AND EXTERNAL MATRIX MULTIPLIERS IN THE INPUT-OUTPUT MODEL .... 59
I.

Introduction ......................................................... 59

II.

Partitioned Matrix Multipliers ....................................... 60

III.

Interregional Repercussion Model ..................................... 65

IV.

Some Extensions of the Model ......................................... 71


1)

Extension in the Number of Partitioned Groups

2)

Inclusion of the Income Formation Process

CHAPTER 5 AN ANALYSIS OF THE INTERDEPENDENCE BETWEEN SERVICE AND GOODS-PRODUCING


SECTORS ............................................................... 76
I.

Introduction ......................................................... 76

II.

Income and Employment Analysis of Interdependency of Two Sectors ..... 77

III.

Input-Output Analysis of the Interdependency of Two Sectors .......... 85


1)

Intersectoral Propagation Pattern

IX

2)
IV.

Cost-Push Effects of Service-Prices

International Comparison ............................................. 93

PART THREE:

DUAL ECONOMIC STRUCTURE

CHAPTER 6 THE DUAL STRUCTURE OF THE JAPANESE ECONOMY AND ITS GROWTH PATTERN
I.
II.

Introducti on ......................................................... 100


Economic Growth and Differentials in Capital Intensity by Size of
Firm ................................................................. 101

III.

1)

Schema of Capital Concentration and Growth of Enterprises

2)

Differentials in Wages and Capital Intensity

3)

Permanence of the Dual Structure

Differentials in Composition of Funds and Interest Rates ............. 111


1)

Funds of Enterprises and Capital Accessibility

2)

Differentials in Interest Rates on Borrowed Funds and Cost of


Funds

3)

Differentials in Interest Rates and Unequal Distribution of


Loans

IV.

V.

Structural Peculiarities of Capital Concentration .................... 120


1)

Factor Proportion and Differentials in Wages and Interest Rates

2)

Structural Peculiarities of Capital Concentration in Japan

Conclusion ........................................................... 126

REFERENCES ....................................................................... 129


INDEX ............................................................................ 133

PART ONE
INPUT-OUTPUT AND INCOME FORMATION
CHAPTER 1
INTERINDUSTRY ANALYSIS AND THE STRUCTURE OF INCOME DISTRIBUTION*

I.

rntnoductio~

In the standard interindustry analysis, consumption demand is treated as an


exogenous variable, so that the usual Leontief matrix multiplier analysis lacks the
multiplier process via the consumption function that one customarily finds in a
Keynesian Model.

In order to treat consumption demand as an

e~0genouo

variable in

the Leontief system, the household sector is routinely transferred to the processing
sectors, and is regarded as an industry whose output is labor and whose inputs are
consumption goods.

But the appropriate procedure in dealing with consumption is not

to regard it as a fictitious production activity, but to introduce the Keynesian


consumptiqn_function on a disaggregated level.

To this end, we have formulated a

matrix multiplier which combines Leontief's propagation process with the Keynesian
propagation process in the form of the Leontief inverse multiplied by a
~nv~e

matnix.

~ubjo~~ed

The subjoined inverse reflects the effect of endogenous changes in

consumption demand. l )
Nevertheless, this extension of the standard Leontief model may not adequately
deal with the interrelation between the interindustry and consumption structures.
The reason for this is' that the consumption structure generally depends on the
structure of income-distribution. The income-distribution structure regulates the
consumption pattern in that the consumption pattern consists of the expenditure behavior of various income-groups.
* This is a revised and integrated version of two articles which are originally
published, under the same title, in M~oeco~o~ca, Vol.15 Fas. 2-3, AgostoDicembre 1963 (with collaboration of Shingo Masegi), and in the theoretical part
of "Input-Output Analysis and Interrelational Income Multiplier as a Matrix,"
H~othub~hi JOUA~at 06 Econo~~, Vol.8, No.2, Feb. 1968.
1) K. Miyazawa [32J, especially Section IV and VI. See Chapter 3 in this volume.

In this chapter, we shall try to incorporate the process of income distribution


and expenditure into the input-output system.

If we denote the income multiplier

manifesting the income-distribution-factors as the "Kalecki multiplier",2) then our


task is one of combining the Leontief output multiplier and the Kalecki multiplier
into its disaggregated and generalized form.

II.

In;tvUndu..6tJr.y Anal.y.6,u, and .the PJc.Oc.e.6.6 06 V,u,:tJc..i..bution and


Expe~e

06 National. Inc.ome

At the outset, in order to delineate the salient aspects our problem, we will
give a brief macro-numerical example of the model to be developed later.

In the

standard input-output model, final demand f (= consumption C + investment I = 10)


determines the level of output X via the input coefficient a = R/X = 3/4 (where R =

total

I
I
I

30
W 6
---------p 4

total X

40

I
I
I
I
I
I
I

intermediate inputs), i.e.

2 40

1
=~
f- = 1 _1 3/4 . 10 = 40. This is a macrocosmic

expression of the Leontief output matrix multiplier.


But consumption C is originally induced by the income Y
( =

wage W + profit

p =

10). The consumption coeffi-

cient is e = e/y = 8/10, so that the Keynesian income


multiplier equation is Y = 1 ~ e

= 1 _ 18/10 . 2 = 10.

Thus by combining the simple Keynesian income multiplier with the simple Leontief
output multiplier, we obtain the following output solution for an input-output model
with endogenous consumption demand: 3)
1
1
1 fX=r:a
=r:-a'~'

I.

(i)

2) M. Kalecki [25], Chap. 5.


3) This macro-multiplier (or its disaggregated form) is derived more convincingly by
tracing the propagation process from the initial injections. This method of
derivation is utilized in Chapter 3.

Of course, the income-multiplier -1_1--holds only for a particular income-dis- a


tribution pattern.

= w/y and

Let d l

profit, respectively, and let a l

d2

= p/Y denote the relative shares of wages and

= Gw/W and

a2

= Gp/P the propensities to consume of

laborers and capitalists, respectively, then we have the generalized Kalecki income
multiplier

~ =1

(ald~

+ a2d 2 ) , which incorporates the income-distribution

factors. 4) The input-output solution is then expressed in the form of (ii):


(i i)

= y/x = 1

If we let v

- a denote the value-added ratio, and v l

= w/X,

v2

= p/x

the value-added ratios of wage-income and profit-income, respectively, the output


solution takes the following form:
(i i i)

This equation (iii) is the macro-counterpart of the matrix multiplier which we will
develop next.

2)

GenVta..Uza;tion 06

the

Inpu.t-Ou.tpu.t Model

The value-added sector in the interindustry model is not only divided into n
industry-groups along the column, but is also divided into r income-groups along the
row, as our simple macro-numerical example illustrates.
of the kth group earned from the jth industry as Ykj (j

4) If we let al

1,0

<

a2

<

Let us express the income

= 1,

... , n ; k

1, we have Kalecki 's own formulation

1
(1 _ dl)(l _ ( ) as a special case of this expression.
2

_-,--'-1-.----,------;-,
1/2, then we have Y = T
- (ald l + a 2d 2 )

-- 1

- 1/2

= 1,

(6/10 + 1/2.4/10) 2
2 = 10.

... , r);

If we let a l
1

= 1,

a2

10 or

this r x n income-formation matrix shows the most generalized pattern of income-distribution.

Corresponding to this income-distribution pattern, consumption demand Cik

is also defined as consumption for the ith commodity by the kth income-group (i = 1,
... , n ; k = 1, ... , r).

The coefficients of our model are represented in Figure 1,


where:

Fi gure 1

A =

the n x n matrix of input coefficients a ..

= x .. Ix.,
1.-J J

V =

the

= Yk

C =

the n x

1.-J

x n matrix of value-added ratios vkj


r

Ix.,

matrix of consumption coeffi cients

a ..

1.-J

x. = jth industry's output,


J

Yk = kth income-group's income.

Let

x = a column vector of output,

fa = a column vector of consumption demand,


i ,j
k

1, 2, ... , n
=1,2, ... , r
(n > r)

f = a column vector of final demand other than consumpti on,

then, the input-output system can be conveniently expressed as

x = AX + fa + f

(1 . 1 )

In the standard input-output analysis where

fa' as well as f, is treated as an

exogenous variable, the following well known solution is obtained:


(1 .2)

But if we treat the consumption demand

fa as an endogenous set of variables and re-

gard the household sector as a distinct decision-making unit instead of as a fictitious production unit, the introduction of a disaggregated consumption function is
necessary.
The consumption function of our model can be written as follows:

= Cv.x = ~l.. ~(k)v(k)x - ~l.. v'k


~ vk .x .
k=l

k=l

1.-

J J

(1 .3)

where

(k) _ (
(k)
- 0lk' 02k' ... , nk) , is a column vector and v
= (v kl ' v k2 ' ... , v kn )

is a row vector.

If we add nonhomogenous terms, or exogenous elements to the con-

sumption function, C becomes the matrix of marginal coefficients, and in this case we
can include the nonhomogenous terms in f. 5 ) Substituting the consumption function
(1.3) into (1.1), we get
(1 .4)

X=AX+CVX+f

Solving (1.4) for X, we obtain the following alternative expressions:


X

= [I

- A - CV]-lf

= B[I

- CVBr If

=B[I

+ CKVB]f

(i)

(i i)

(1. 5)

(i ii)

where, of course, B = [I - A]-l, i.e., Leontief inverse matrix multiplier.


The first expression (i) in (1.5) gives us the
pli~

~ged inv~e

matnix multi-

showing the total effects of exogenous final demand on outputs via interindus-

try and induced consumption activities. The existence of the inverse [I - A - cv]-l
is generally verified.

Expression (i) can be converted into the second expression

(ii), namely, the "original Leontief inverse"

postmultiplied by the inverse

The conversion is as follows:


[I

- A - cv] -1

[{I - CV(I - A)-l}(I - A)]-l


(I - A)-l[I _ CVB]-l

= B[I
We can refer to the inverse

(1 .6)

- CVBr l

[I - CVB]-l

as the hubjoined inv~e matnix.

This in-

verse reflects the effect of endogenous changes in each income-group's consumption


expenditure.

Matrix multiplier equation (1.6) corresponds perfectly to the macro-

5) If we define some o(k), which is the capitalist group's coefficient, as <propensity to consume ptu6 propensity to invest>, our model formally contains the problem of induced investment.

multiplier (iii) in our previous example. 6)


The advantage of matrix multiplier formula (ii) in (1.5) is that it distinguishes the inverse reflecting endogenous consumption activity from the inverse reflecting production activity, in contrast to formula (i) which does not make such a
distinction.

Moreover, if consumption coefficients and value-added ratios are not as

stable as the input coefficients, it is desirable to have the "subjoined inverse"


expressed in a form which can be easily computed and revised.

The development of

such a practical computation formula is also useful from the standpoint of understanding the theoretical aspects of inter-income group activity.

We now turn to such

a task.
Let us write
k=l, ... , r
v=l, ... ,r.
Then, as we shall show, we can prove that:
B[I -

CVBr 1

= B[I

(1.7)

+ CXVB].

The third expression (iii) in (1.5) means that the n x n subjoined inverse

[I - cv.s]-l can be obtained, without inversing the matrix, by the means of using the
~~etational

income

muttipli~

X whose order is r x r.

A proof of the identity

between (ii) and (iii) is as follows:

=I

with the definition

X[I - VBC]

then

CX[I - VBC]VB

'"

= CVB

CXVB[I - CVB] =CVB


I - CKVB[I - CVB]
I

= [I

=I

- CVB

CXVB][I - CVB]

:. [I - CVBr l

=I

+ CKVB,

where identity matrices I's in the first and second equations have the order of
6) If we set r = 1 in (1.6), i.e., if we do not make a distinction among the incomegroups, the equation (1.6) coincides with the formula which we have derived elsewhere (see [32] p. 63 or (3.20) in Chap. 3), and it corresponds perfectly to the
macro-multiplier (i) in Section 1).

7
l' X 1',

and those in the third and subsequent equations have the order of n x n

respectively.
the

l'

l'

In practical terms, since

l'

in most cases is very much smaller than n,

matrix K should be readily obtainable.

Consequently, if we already have

the numerical table for B, we can renew the subjoined inverse whenever it is necessary to do so.7)

3)

The Coe6McLent6 06 IntVt- Inc.ome-GlWupb

We may also work out the proof of formula (iii) in (1.5) by the method which
traces the propagation process initiated by the original injections.

This method

may, at the same time, reveal the economic meaning of matrices Land K.
Denoting by m the numerical stage of the propagation process, we get

(m 1; 2)

(1.8)

Hence,
(1. 9)

Thus,

x=

m=l m

Hence, if the term


X

= B[I

= Bf

+ BC(

Lm- 2 )VBf.

m=2

(1.10)

I Lm- 2(i.e. m=OI Lm) is convergent,8)

m=2

+ C(I - L)-lVB]f

= B[I + CKVB]f.

(loll)

The result again confirms (1.7).


The matrix L

= VBC

may be interpreted as an array of coefficients which show

the interrelationship among income-groups in the process of propagation resulting


from each income-group's consumption expenditure pattern.

In order to prove this

7) Our model can be easily extended to accomodate an open economy with foreign trade.
8) The convergence conditions of our model will be examined in Section IV.

point, we take the vth income group as representative and trace its consumption expenditure effect on another kth income group's income.

increase in output
of each industry

->

increase in income

->

of the vth group

increase in output of each

->

increase in consumption

->

of the vth group

increase in income of the kth

industry due to the additional

group due to additional income

consumption of the vth group

of the vth income group

Thus the element of L, i.e. Zkv' can be written as


(1.12)

That is, the coefficient Zkv shows how much income of the kth'income-group is generated by the expenditure from 1 unit of additional income of the vth income-group.
Thus we can term L the "matrix of inter-income-group coefficients", and K "the
interrelational multiplier of income groups".
A proposition arises in connection with the matrix of inter-income-group coefficients:

the column sums of the matrix

equal the total consumption coefficients

of each income group, i.e.


i'I' L

where

i~

= i'VBC
= i'[I
- A]BC = i'C
I'
n
n'

and i; are row-summation vectors of order n and

III.

The. Re1.a.UonolUp

(1.13)
I'

respectively.

06 In.:teJt-Income.-GJtouP.6 and

the. Mutti-Se.etoJt Income. MuttiplieJt

1)

The Inc.ome MuLUpUeJl.

a6

a. Ma.tJvi.x

We shall now project equilibrium output into equilibrium income.

As before,

denoting by Y the column vector of r order whose elements are household incomes by
income-groups, we get
(1.14)

Y = VX.

substituting formula (iii) of (1.5) into this expression (1.14), the income equation
becomes
Y = VB[I + CXVB]f

[I + VECX]VEf

(1.15)

[I + LX]VBf,

in which r + LX

= X because [r -

L]K

= I, so we obtain
(1.16)

Y = KVBf.

Justification for the existence of formula (1.16) may be attempted by tracing


the propagation process caused by the initial autonomous injection of f, or final
demand excluding endogenous consumption expenditure.

Using suffix m in parentheses

( ) to denote the numerical stage of propagation, we get


VX{l)

= VBf
(1.17)

VX (m) = VECY (m- 1 )


= LY{m-1)

m-1

=L

Y{l)'

for m

This gives the expansion in powers as:


00

Y =
=

m~,r (m)

= Y (1)

[r + L + L

+ LY (1) + L Y (1) + L Y (1) +

3
+ L + ... ]Y (1 ) .

(1.18)

Hence, if the term Lm is convergent, we obtain the following fundamental equation of


income formation:
Y

= [I

- L]-l VEf

(1.19)

= KVBf

We may des i gnate the r x n matri x KVB as the


matrix form or simply ma.tJvi.x muLUp..{eJt

on

muLU-~ec.:toJt

..[nc.ome muLUpUeJl. in

..[nc.ome noJtma.t..[on.

This matrix has the

10
following composition:

the "interrelational income multiplier"

the coefficient matrix of induced income

VB. 9 )

post-multiplied by

Thus, equation (1.19) will give us

the direct and indirect induced incomes of each income-group attributable to the
initial autonomous demand. 10 )
This multi-sector income multiplier is a distinguishing feature of our model.
In the conventional input-output analysis, where consumption demand is entirely
exogenous, the outputs of various industries have different values depending on the
proportions of final demand;

but as far as the value-added sector is concerned, in-

come has the same value as final demand and does not depend on the proportions of
final demand.

In contrast, as is evident in (1.19) of our model, incomes (both total

income and group incomes) have different values depending on the proportions of final
demand, and this is due to the fact that our model takes explicitly into account the
structure of income distribution.

2)

Ac.c.epted Mu1:UpUeM M Special.

eMU

This conclusion cannot be obtained by the introduction of an endogenous consumption structure without some explicit consideration of the distribution-pattern.

The

reason for this is as follows.


(a)
the matrix

If we do not distinguish among the income-groups, i.e. if we let r = 1,


V

becomes the row vector of n order and, correspondingly, matrix

comes the column vector of n order.

be-

If we denote these vectors as v' and a,

respectively, and assume that all value-added in the national economy consists of
the income accruing to the household sector,ll) then
9) An alternative justification for formula (1.19) was suggested by W. H. Branson
[4] at the Econometric Society Meetings, Washington D.C., 1967. Income generated by exogenous expenditure is equal to VBf, and income generated through
endogenous demand as a function of income is equal to VBCY, thus income Y is
given by Y = VBCY + VBf = [I - VBCr 1VBf.
10) To combine the income-effect in our model with the relative price-effect, we may
be utilized R. Stone's "linear expenditure system". See [50], [52].
11) With this assumption, v' becomes the vector of value-added ratios for the whole
economy, and in an economy with no foreign trade and government activities, the
conversion v' = i'[I - A] becomes possible. Then we get
v'B = i'[I - A][I - A]-l = i'I = i'.
Of course, if the household sector accounts for only one part of the value-added
sectors in the national economy, this conclusion must be modified.

11

L = VBC = v'Ba = i'[I - A]Ba = i'a

= a (=Keynesian macro-propensity to consume)


-1 = 1K = [ I - L]

(1. 20)

1 - 0

where i' is a row-summation vector.

So, the income multiplier equation (1.19) be-

comes
Y = KVBf

= _1_ v'Bf = _1_ i'f = _1_ f


1-0
1-0
1-0 a

(1.21 )

where fa is a scalar, where fa = fl + f2 + f3 + ... + fn' and the vector Y becomes a


scalar, too.

This scalar multiplier coincides exactly with the Keynesian multiplier.

Thus our conclusion that income has different values depending on the proportions of
exogenous demand is not substantiated in the special Keynesian case. 12 )
(b)

Furthermore, even if we introduce income-distribution-factors in macro-

economic form as in the Kalecki or Kaldor models, the above Keynesian result is not
improved.

Denoting by

the column vector of

order whose elements are relative

shares of each income-group,13) we may rewrite the matrix Vas

= dv', and the

matrix L takes the following form:


L

= VBC = dv'BC = di'C = de',

where e' = i'C is the row vector of k order whose elements are the total propensities
to consume of each income-group. Then, we get Lm = (de,)m = d(e'd)m-l e , = dim-le',
where i is a scalar showing the weighted average of propensities to consume of each
income-group. Thus, the interrelational income multiplier in this case is
K

= [I =I

L]-l

=I

I i m- 1de' =

m=l

I Lm

m=l

I + _1_ de' ,

1- i

12) The output multiplier corresponding to this case will be


X = B[I + _1_ ai']f
1 -

and this coincides with the result (3.20) in Chapter 3.


13) Where, of course, the sum of all elements of d is equal to 1, i.e.,
12k _
d+d+ ... +d-1.

(1 .22)

12
and the fundamental equation takes the form:
Y

= KVBf
+ _1__ de'] dv'Bf = [I + _1__ de'] di'f = [d +

= [I

1-1.

1-1.

1-1.

(1. 23)

= _l-df.

1 -

~d]fO

In which case, the autonomous demand vector f becomes a scalar fO' and the equation
(1.23) coincides with the Kalecki multiplier, except when it is expressed in some
generalized form.

In order to convert the above equation into a scalar multiplier,

all that is required is to multiply both sides of the equation by summation vector
i', i.e.,
i'Y

= i,_l_ df = i'd-ll-i

l-i

= _1_ f
l_i

(1. 24)

If we assume the constancy of relative shares, the scalar 'i always takes a constant
value, and, after all, equation (1.24) ends up being formally equivalent to the
Keynesian multiplier (1.21).
(c)
customary

Again, if we regard consumption demand as an exogenous variable as is

(f
Y

= f + f), the income multiplier equation becomes


()

= VBf = v'Bf = i'f = fO'

(1. 25)

and income equals final demand irrespective of the proportions of final demand.
Thus, in order to conclude that the values of income differ depending on the
proportions of autonomous final demand, it is necessary to introduce not only the
structure of consumption demand, but also the structure of income distribution.

3)

Sbw.c.twr.e 06

.the PILOPll9ilioYl PILOC.e1>.6

If we lump together the above two mechanisms of output and income determination,
we have the following system:

~]

[*J [~]

~].

Solving this system for X and Y, we get

(1. 26)

13
and it is expected that this solution can be converted to the form:
(1.27)
where g is a column vector of exogenous income. 14 ) The preceding separate solutions
(1.5) and (1.16) are equivalent to (1.27) where g is disregarded.
Now, let us return to the output propagation equation (1.10).

Equation (1.10)

can be interpreted as the propagation process viewed from the income-formation side.
But the same propagation process can also be observed from the consumption side or
the production side as well.
(a)

the income-formation side (VBC


x

= Bf
=

(b)

These three aspects of the propagation process are:

= L)

+ BC[I + VBC + (VBC)2 + ... ]VBf

Bf + BC[I - L]-lVBf

(1 .28)

the consumption expenditure side (CVB)


x

= B[I

+ CVB + (CVB)2 + ... ]f

= B[I + C(I - L)-lVB]f


(c)

(1. 29)

the production side (BCV)


X = [I + BCV + (BCv)2 + .. ]Bf

(1. 30)

[I + BC(I - L)-lV]Bf.

It is interesting to note that in all cases, we can obtain the computation formula (1.7) by projecting the propagation process into the income-formation side
VBC.

On the other hand, if we derive the sum of the geometrical progression from the

consumption side (CVB) or the production side (BCV), we do not obtain the computation
formula (1.7) directly, but instead obtain the equation (1.6) which is the product of
two inverse matrices.

This means that the income-formation side has a homogeneous

14) The proof of (1.27) is easily demonstrable by use of the following identity:

IPEI + CKVB] IBCK] [r - A I-C]


L KVB
IK L-V 0

WQ]
iPrI '

The expression (1.27) in this chapter is equivalent to the formula in K. ~1iyazawa


[33], or (4.7) in Chap. 4, in which, if we let Bl = VB, B2 = BC and M = K, we get
(1. 27) .

14

character which contrasts strikingly with the nonhomogeneous character of both production and consumption activities.
One other point regarding the propagation process should be explained.
tions (1.28)

Equa-

(1.30) assume a propagation process in which the entire process is a

succession of separate two-step movements:

in the first, the propagation from the

production side is represented entirely by the effect of matrix B, and in the next
step, the propagation occurs on the income-formation and consumption expenditure
sides.

But instead of this assumption, we may assume that propagation occurs simul-

taneously in all three sides, i.e., production, distribution and expenditure.


latter case, instead of equation (1.28)

In the

(1.30), the propagation equation may be

rewritten as follows:

x =f
We write

A +

(A

CV)f

(A

Cv)2f +

(1.31 )

CV = Q, and, if we assume the term

00

L~

m=O

to be convergent, we have
(l .32)

which coincides with (i) of (1.5).


X

= B[I

- CVB]-lf

= B[I

By formulae (ii) and (iii) of (1.5), we get


+

CKVBJf.

Thus, the two propagation cases, i.e. the case of (1.28)


(1.31), have the same sum, but obviously the
(1.28)

~ncated

(1.30) and the case of

muttiplien in the case of

(1.30) has generally a larger value than the truncated multiplier in the

case of (1.31).

We turn next to the analysis of the convergence conditions of these

two cases.

IV.

The Convengence CondLtiOn6 06 the Model

So far, we have assumed the existence of a meaningful solution, X


fundamental equation, X
(I - A - CV) -1

= AX

(I - Q) -1

+ CVX +
~

f, (f

0, for our

0), i.e., we assumed the existence of

1.
0 and of K = (
I - L)

In order to treat these prob-

lems and their relationships, we will first review the properties of the Leontieftype matrices as preparation for developing the convergence conditions of our model.

15
1)

The PfWpeJLtie.-6 06 Leon.tie6-type Ma:tJUce.-6

For non-negative square matrices in general, the following properties are well
known:
[I] Let a be a n x n non-negative matrix.

Then the conditions (1) - (4)

below are equivalent.


00

(1 0)

L am

m=O

converges

(2)

All characteristic roots of a are less than 1 in absolute value.

(3)

I - a is non-singular and (I - a)-l is non-negative

(4)

For any non-negative vector f, the equation (I - a)x

= f has a

unique non-negative solution.


For Leontief-type matrices,. i.e. non-negative matrices with no column-sums
greater than 1, Woodbury gives the following lemma: 15 )
[II]

(I -

a)x

Let a be Leontief-type and

nonsingular.

Then the equation

= f has a unique non-negative solution.

From propositions [I] and [II] we obtain


00

Lemma 1.

Let a be Leontief-type.

L am

m=O

converges if and only if I - a is

nonsingular.
Now, we may transform a into the form (1.33) below by some permutation matrix

(1 .33)

A2 . A2k

o
where Al , A2 , ... , Ak are indecomposable square submatrices, and k

>

2 or k

=1

depending on whether or not a is decomposable.


Then we may improve upon another proposition of Woodbury's.16)

15) M. A. Woodbury [54], p. 353, Lemma 3.2.


16) M. A. Woodbury [54], p. 357, ConoLt~y 3.6, where the condition is stated as
follows: "at least one of the column sums be less than 1 for some column in
each block. 06 columlU> of the matri x" .

16
[III] Let a be Leontief-type.
any non-negative

vecto~

t,

A necessary and sufficient condition that

the equation (I - a)x

=t

ha6 a non-negative

6o~

~o~on

is that at least one of the column sums be less than 1 for some column in each
submatrices Al , A2 , ... , Ak in (1.33).
Based on [I], another form of [III] is obtained by replacing the paragraph
italicized in [III] with "all characteristic roots of a be less than 1" in absolute
value, which we call [III'].
Solow's Theorem asserts that condition [III'] is a sufficient one. 17 )
We can also show that it is necessary too. 18 )
A different form of [III] or [III'], more convenient for our purpose, is
Lemma 2. 19)

Let a be Leontief-type.
00

(1)

If all column-sums are less than 1,

(2)

If all column-sums are equal to 1,

L am

m=Q

converges.

00

L am

m=Q

diverges.

17) R. Solow [47], p. 36, Theorem 1 and p. 38, Corollary.


18) Proof of the necessity of condition [III'].
We show that (i) implies (ii) below.
(i) All characteristic roots of a are less than 1 in absolute value.
(ii) Each Al , A2 , ... , Ak has at least one column-sum less than 1.
Suppose that the condition (ii) does not hold. Then all column-sums of some
A. are equal to 1. For m-dimensional vector j = (1, 1, ... ,1), m being the
d~gree of A., jA. = l.j, i. e. 1 is a characteri stic root of A.. As the
'I.
'I.
'!.
characteristic roots of Al , A2 , ... , Ak are also that of a, 1 is a characteristic root of a, unlike (i).
19) Proof of Lemma 2.
(1), (2) and the necessity of condition (3) are immediately evident from
[III'] and [I]. (As to (1) and (2), see also R. Solow [47], p. 32, p. 37).
00

L am diverges. Then, by [III'],


m=Q
a must be decomposable and, for some A. in (1.33), all column-sums are equal to
1. If i = 1, let Al = A(l). If it- 1: Al ., ... , A. 1 . are all zeromatrices
'I.
'1.- ,'I.
(otherwise, at least one column-sum in the i-th block of columns of a must be
greater than 1). Hence we can remove A in (1.23) to the upper left corner by
some simultaneous permutation of rows and columns, without losing the character
of the form (1.33).
Then we let A.'I. = An (1).
In either case, a is decomposable into the form (1.34).
Sufficiency of condition (3):

Suppose that

17

(3)

In case some column-sums are equal to 1 and some less than 1,


00

L am

converges if and only if a is not decomposable into the

following form (by some simultaneous permutation of rows and


columns) :

~(l)

where all column-sums of A(l)

A(l2) ]
A(2)

are equal to 1.

(1. 34)

The assertions (1) and (2) in this proposition have nothing to do with the
And as for (3), it is the particular and not the general

decomposability of a.

decomposability which matters.


To be precise, condition (3) includes two cases:

the case where a is

indecomposable and the case where a is decomposable, but not into the form (1.34).
Now, let us return to our model.

We may assume that the matrices A, V and e

in the preceding sections have the following properties [pl] - [p4].


[pl]
[p2]
[p3]

l'

L a.. + Lvk

i=l

k=l

1.-J

l'

Lv
k=l kJ

>

0 or

>

LVkj

j=l

=1
n

La1.-J..

i=l

<

(j

1, 2, ... , n)

(j

1, 2, ... , n)

(k

1, 2, 00 ., 1')

(k = 1, 2, ... , 1'), where a k

[p4]

These assumptions are reasonable from an economic standpoint.

(A generaliza-

tion of [pl] is to be examined later).


The existence of B = (I = A)-l is guaranteed by [p2J (See Lemma 2 (1)).
A, V and e are respectively non-negative, n x n,
=

l'

x nand n x 1', matrices, VEe = L

(Zk v ) and A + ev = Q = (q 1.-J


.. ) are also respectively non-negative,

l'

matrices, and the following equalities hold:


Lemma 3.

00)

l'

L Zkv =aV

k=l

As

(v

1,2,00',1')

l'

and n x n,

18

(2)

i=l

1,2, ... ,n).

q .. = 1
1-J

From the equalities and [p4]:


Land Q are Leontief-type.

Co~o~y.

2)

Convengenee

Con~On6 ~n ~he

Model

We can now consider the convergence properties of the propagation process in our
model.

= jn. Thus,
+L

Ign

'"
I~.

m=O

is also Leontief-type.

m-l) VB, the convergency of

Lemma 1,

'" m
I L coincides with that of

The convergency of

Theo~em 1.

m=O

Since I + R .. + ~ = I + C(I + L +

-m
IEm
and that
of I~

and IQm converge if and only if

are equivalent.

I - R

f 0 and

Next, from
I - Q

f 0,

respectively.
And, since I where always

I - A

= I - A - CV = (I - CVB) (I - A),
f O.

Hence

I - Q

f 0 and

I - Q

I - R

I - R

I I

I - A

I,

f 0 are equivalent.

This means that I~ converges if and only if I~ converges, and therefore if and only
if

ILm converges.
A simultaneous permutation of rows and columns in A, reflecting a change in the

order of industry groups, induces a permutation of the columns in V and that of the
rows in C.

On the other hand, a permutation of rows in V, reflecting a change in

the order of income groups, induces a permutation of columns in C, and conversely.


For brevity, we call the former I-permutation and the latter II-permutation.
Then, as a convergence condition of I~, we have:
Theo~em

2.

Let ak =

i=l

(k

a. k

1, ... , X')

1-

be the total propensities to consume of income groups.


(1 0)

If all a l , ... ,

aX'

are 1ess than 1, then I~ converges.

19

1,

LQF

= 0p = 1, then LQF diverges.

(2)

If 01 = ...

(3)

In case some of ok (k = 1, ... , p) are equal to 1 and some less than

converges if and only if A, V and C are not decomposable by any 11- and

I-permutations sumultaneously into the following respective forms:

c=

(1.35 )

V=

where 0

<

Proof.

<

n, 0

< 8 < P,

and all column-sums of Cl are equal to 1.

From Lemma. 3, (2),

L q ..

i= 1 1.J

(k
(1)

ok

L q1.J.

i=l

= 1.

L q .

i=l

1.J

<

<

1 (k = 1, ... , p):

= 1,

= 1 if and only if
... , p)

(i )

Suppose that there exists a number j such that

Then v l . = .. = v . = 0 from (i).


J
PJ

1 for all j, and therefore

LQF

This contradicts [p2].

Hence,

converges (Lemma. 2, 1).


n

(2) 01 = ... = o~ = 1: As equation (i) holds for all j, L q .. = 1 (j = 1, ... ,


"
i=l 1.J
Thus,

LQF

n~

diverges from Lemma. 2 (2).

(3) A simultaneous permutation of rows and columns in Q induces a I-permutation of


A, V and C only, and conversely.

A II-permutation leaves Q unchanged.

Therefore

LQF

converges, from Lemma. 2. (3), if and only if Q is not decomposable by any I-permutation of A, V and C, into the following form:

20

where all column-sums of Q1 are


equal to 1, 0

<

<

(i i)

n.

We shall prove that Q is decomposable into (ii) if and only if A, V and Care
decomposable into the respective forms represented in (1.35).
Let Q be decomposed into (ii).

By a suitable II-permutation, without changing

Q, we may take
(i i i)

(O<t<1')

As all column-sums of Q1 are equal to 1, we have, from (i) and (iii),


Vkj

(k = t + 1, ... , 1';

= 0

Again, from the form (ii),-a .. +


'l-J

(i = h + 1, ... , n;

l'

L c.kv k .

k=l

(i v)

j = 1, ... , h)

'I-

=0

1, ... , h).

As all terms on the left-hand sides of these equations are non-negative,

=0

(i = h + 1 ,

... ,

n;

= 1,

. .. ,

cikvkj = 0

(i = h + 1 ,

... ,

n;

= 1,

... , h;

a .
'l-J

(v)

h)

k = 1,

... , 1')

(vi)

We may take, if necessary, after transforming by some II-permutation, the first


a rows of

such that vk.

Jk

>

0, for k

= 1,

... , B and for some jk (1 ~ jk ~ h), and

obtai n

=0

V kj

Here 0

< B ~

<

(k = B + 1 ,

... ,

n, from [p2] and (i v) .

cik = 0

(i = h + 1 ,

t;

= 1,

... ,

h)

(vi) ,

Then, from (vi) ,

... , n;

= 1,

... ,

B)

(vi i )

The last II-permutation changes the order of columns in c, but only among the
first

columns.

Hence (iii) remains valid.

And, as

is unchanged, so is (v).

From (iv), (iv'), (v) and (vii), we conclude that A, V and C are decomposable into
the forms depicted in (1.35).
The converse is obvious, and the proof is completed.

21
The condition in (3) of this theorem contains the following two cases:
(a)

is indecomposable, (b)

is decomposable, but not into (1.35) together with

and C simultaneously.
As a consequence we have
Corollary 1.

In the case where some c l ' ... , cr are equal to 1 and some less
than 1, L~ comverges if

is indecomposable.

So we can treat Kalecki 's Model as a special case of

ConoLt~y

1, where

n = 1, r = 2.

According to Theonem 1, we have


ConoLt~y 2.

The convergence condition of L~ given in Theonem 2 is also that


of LLm.

In addition, from [I] we have


Cono~y

3.

A necessary and sufficient condition that, for any f

0, the

equation (I - Q)X = f has a non-negative solution coincides with


the convergence condition given in Theonem 2.
The zero parts of A, V and C in (1.35) may be as large, but no larger as V2
and Cl will vanish in view [p3] and [p4].
Moreover, we may adopt a more general assumption than [pl]:
[pl' ]

La .. + L V k

i=l 7-J

k=l

~ 1.

Then, the possibility of L~ diverging becomes more limited:


remains true, but (2) no longer holds.

(1) in Theonem 2

As for (3), the given condition is a suf-

ficient, but not a necessary one. 20 )

20) One case represented by assumption [pl '] is an open economy with foreign trade.
As shown in Section III, in the case of a cio~ed economy with no foreign trade,
the conclusion that income has different values depending on the proportion of
final demand can be derived only by introducing the structure of income distribution. But in the case of an open economy with foreign trade the same conclusion can be derived without introducing the distribution structure, because the
coefficients on the production side enter into the income-formation process
through imports. For the open economy model, refer to Chapter 3.

CHAPTER 2
INPUT-OUTPUT ANALYSIS AND INTERRELATIONAL INCOME MULTIPLIER AS A MATRIX*

I.

rntnoduction

As an extension of the input-output analysis, we have introduced in the previous


chapter a new concept which might be called the "interrelational income multiplier"
in matrix form.

It was designed to analyse the interrelationships among various

income-groups in the process of income formation, and in this respect it tells us how
much of one group's income is generated by another group's expenditure from one unit
of additional income via the medium of industrial production activity.

Although this

multi-sector multiplier follows from Leontief's "interindustry matrix multiplier", it


is formulated by the inclusion of the income generation process, which is omitted in
the standard input-output open model, and by projecting the multiplier process into
the income-determination side rather than the output-determination side.
Our extended model contains a theoretical implication not found in the Keynesian
model nor in the standard Leontief input-output model.

In the Keynesian income-

determination multiplier model, the same amount of autonomous expenditures cannot


have different effects on the level of national income, even though the expenditures
have different commodity proportions.

The same restriction holds for extended models

incorporating income-distribution-factors, such as the Kalecki-Kaldor type, in so far


as there are no changes in the relative shares of income and in the propensities to
consume of each income group.

Similarly, in the Leontief input-output model, al-

though the outputs of industries vary depending on the proportions of autonomous


expenditures, the total income is independent of the composition of autonomous expenditure. This result also holds in the case where household consumption expenditure is treated as an endogenous variable, in so far as we retain the assumption that

* This is a slightly revised version of the empirical application part of article


which was originally published, under the same title, in Hita~uba6hi ]o~nat 06
Eeanomic6, Vol. 8 No.2, Feb. 1968.

23

the level of income and its use do not depend on the composition of production. l ) In
the real world. however. autonomous expenditures of equal amount. but having different commodity compositions. appear to have different effects on income formation.

In

order to have an input-output model in which the value of income differs depending on
the proportions of autonomous demand. it is necessary to introduce not only endogenous consumption in disaggregated form. but also the structure of income distribution
by the type of income-group as well as by the tupe of industrial value-added.

This

is exactly what our extended model accomplishes.


In this chapter we shall see whether or not our extended model is consistent
with facts. We shall use the model to interpret input-output and related empirical
data.

II.

Fotunula.tion 06 a Regional. Inc-arne !Au.;t[p'ueJt in ;the 60tun 06 a MabUx

An application of our model is made for interregional income-distribution.


model is thus reformulated in a form suitable for this purpose.

Our

However. the gener-

al model itself is applicable to the study of class-distribution or size-distribution of incomes as well.

The ommision of the income formation process in input-

output analysis is especially unwarranted in the interregional interindustry case.


because the location of production depends on the location of consumption. and the
latter cannot be determined separately from the calculation of income generated in
each region.
Let us divide k regions into n industry sectors. and express the coefficients
of the model as follows:
A =

[{jJ ...
a~~
1.-J

the nk x nk matrix of interregional input coefficients.

the amount of ith commodity produced in region r for use of 1


unit of output of the jth industry in region s (i. j
.... n;

= 1. 2

r,s=1.2, ... ,k).

1) This is true in the case of a closed economy with no foreign trade and government
activities. But in the case of an open economy with foreign trade and government
activities, the same conclusion does not hold, because the composition of production plays a part in the income formation process through imports, subsidies and
taxes.
.

24

... the k x nk matrix of value-added ratios of household sectors

[y~SJ
J

in each region.
the income of a household in region r earned from 1 unit of

v~S
J

production of the jth industry in region

(j

1,2, ... ,

n;

r,s=1,2, ... ,k).


the nk x k matrix of coefficients of regional consumption
expenditure.
rs

the consumption expenditure for the ith commodity produced in

ci

region r from 1 unit of income earned in s region's household


sector (i = 1,2, ... , n; r, s = 1,2, ... , k).2)
Using the nk x nk matrix of the usual Leontief interindustry inverse:

r1=

B = [I - A

[b~~J,
1.-J

(2.1)

the corresponding earnings by the household sectors in each region are easily determined as follows:
VB=

\' rs sr]
.. ,
~ lv.b
,s

(2.2)

J1.-

which forms the k x nk matrix of coefficients showing induced income earned from
production activities among industries and regions.

On the other hand, the induced

production due to endogenoU4 consumption per 1 unit of income in each region's household sector is given as the following nk x k matrix:
Be =

rLz:,rI b~~Cx:sJ.

(2.3)

J1.- 1.-

Joining these two expressions, we get the following square matrix:

(2.4)
2) Instead of this definition, if we denote byeS the total propensity to consume in
region s, and by h~s the consumption-allocation coefficients for the ith commodi1.-

ty produced in region r, then we get eSh~s


1.-

= cX:1.- s

25

since the multiplication of rectangular and square matrices V, Band C makes a new
k x k square matrix whose order is equal to the number of regions.

This square

matrix L may be interpreted as an array of coefficinets which shows the interrelationships among incomes of various regions through the process of propagation from
consumption expenditure in each region.

Its elements Zrs show how much income in

region r is generaged by the expenditure from 1 unit of additional income in the


region s.
Of course, the income propagation process does not terminate after the one
round of inducement indicated by the coefficient matrix L, because the next round of
earnings will be generated as a result of the production activity induced by the expenditure from the preceding round of additional incomes.
as the "successive income generating process".

This has come to be known

Ultimately, such a successive reper-

cussion process naturally leads to the intersectoral income multiplier among regions
of the following type:
K

The matrix

[I - L

r 1 = [krsJ.

(2. 5 )

will be called the "interrelational regional income multiplier matrix".

This matrix shows the direct and indirect income-generaged per unit of income regionally originated, where, of course,

is the identity matrix having the order of

k x k.
Denote by X a column vector of nk order whose elements are output of n industries in each region, and by Y a column vector of k order whose elements are incomes
of the household sector in k regions.

Let f stands for a column vector having nk

order of final demand other than endogenous consumption expenditure, and g stands
for a column vector having k order of exogenous income. Then we have the following
system:
(2.6)

As shown in Chapter 1, solving this system for X and Y, we get

[~] = ~[I ;~KVBJ IBiKJ ~J.

(2.7)

26
III.

The

Int~elational

Income

Muttipti~

among

Reg~on4

An empirical application of our model is made for a three-region view of the


Japanese economy by utilizing the large 1960 interregional input-output table published in 1966 by MITI (the Ministry of International Trade and Industry). The table
required more than three years of preparation prior to its publication.

For analyti-

cal purposes, the original data tabulated in 9 blocks and 25 industrial sectors was
aggregated into three regions and 25 sectors.

All estimations of parameters and

other calculations for our model were done under the cooperation of the MITI-staff,
together with some other calculations for general or conventional studies. 3)
The logic of separating the Japanese economy into three regions, namely, the
Northeast, Middle and West, is based mainly on our desire to study the interrelationships between the advanced and backward areas in Japan. 4 ) Of these three regions,
the Northeast and West are relatively backward, and the Middle is the advanced area
in Japan.

The major industries are concentrated in the Middle, except for special-

ized regional sectors which exist because of resource endowments and other reasons.
Table 1 is the interrelational income multiplier for the three-region model.
The equations used here have been given in (2.4) and (2.5), i.e., K = [I - LJ- l =
[I - VBCJ- l , where the orders of matrices V, Band C in this case are 3 x 75,75 x 75

and 75 x 3 respectively.
endogenous.

We take only the consumption of the household sector as

Thus, the elements of V are not the value-added ratios for the economy

as a whole, but merely that of the household sector with the fraction corresponding
to the retained income of the business sector excluded.

The estimated parameters for

3) See, Ministry of International Trade and Industry [24J. The details on dataarrangements for application of our model are shown in Chapter 2 of the Supplement Part of this report.
4) The Northeast is defined to include Hokkaido and Tohoku, the Middle includes
Kanto, Chubu and Kinki, and the West consists of Chugoku, Shikoku and Kyushu. The
manner in which Japan is divided into these blocks is based on the Regional Bureaus of MITI for the convenience of collecting regional data. Therefore, aggregation into three regions, consistent with economic areas, is somewhat restricted
by this base.

27
the coefficient matrices V and C are given in the Appendix tables in summary form. 5)
TABLE 1
INTERRELATIONAL INCOME MULTIPLIER FOR THE THREE-REGION MODEL
K = [I - VBCr 1

region
inco~e.
of income orlgln
receipt

Northeast

Middle

West

Total

Northeast

1.55

0.07

0.04

1.66

Middle

0.32

1.57

0.29

2.18

West

0.06

0.07

1.60

1.73

1.93

1.71

1.93

5.57

Total

The interpretation of the entries in the interrelational income multiplier


presented in Table 1 follow from the nature of equation (2.5). The (p, s)th entry
represents the total household income of the pth region induced by expenditure from
1 unit of income earned in the sth region.
column of the Table tell us that:

For example, the figures in the first

1 unit of income earned in the Northeast gives

rise directly and indirectly to 1.55 units of income in the Northeast itself, 0.32
units in the Miadle, 0.06 units in the West, and, in total, 1.93 units for the
national economy as a whole through the medium of expenditure and production
activities.
Let us now compare the column totals in the last line of the Table. These
5) We divide the consumption demand of the household sector into endogenous and
exogenous items, and only the former is included in the multiplier side. The ex~
ogenous household consumption included in the multiplicand side consists of the
expenditure from the transfer income from government, and its propensity to
consume is assumed equal to one.
With regards to thousehold income, we take the compensation of employees, wage
income of unincorporated enterprises and income from properties as endogenous
items,while the retirement allowance in the private sectors and transfer income
from government are treated as exogenous elements. The estimations of income and
its allocation among regions and sectors are mainly based on the income statistics
and input-output data.

28
values show the induced effects of each region's income on the household incomes of
the national economy as a whole.

Among these column sum values, the figure for the

Middle, 1.71, takes a slightly lower value than those for the Northeast and West
(both equal to 1.93).

This is mainly because:

(1) as shown by the coefficient

values of matrix C in Table b of the Appendix, the propensities to consume is


relatively low in the Middle as compared with the Northeast and West, (2) as shown by
the coefficient values of matrix V presented in Table a of the Appendix, the weight
of manufacturing sectors which have low value-added ratios is relatively large in the
Middle area, and finally (3) the weight of the primary sectors which have high
value-added ratios is relatively large in the Northeast and West.

However,

practically speaking, these regional differences in column totals are of little


significance.
Against these column sum values showing the induced effects

o~gin~ng

from

each region, the values of row totals on the right hand side of the Table show the
induced effects

~eceived

the regions of origin.

in each region due to expenditure from 1 unit of income in


In contrast to the column sum values, there are large

differences in the row sum values.

The Middle area receives induced effects

amounting to 2.18, while the Northeast and West receive effects amounting to only
1.66 and 1.73 respectively.

The contrast afforded by the row sum and column sum

values of the interre1ationa1 income multiplier matrix shows a characteristic feature


of the Japanese interregional economy, namely, the concentration of income formation
in the advanced area.
To understand this point more clearly, let us look at each element in the matrix
K.

As shown in the column elements of the Middle area, the induced effects of the

Middle on each region appear intensively in the Middle itself, which takes a value
of 1.57, and appear only negligibly in the Northeast and West, where the values are
identically equal to 0.07.

By contrast, the induced effects of the Northeast and

the West on the Middle take the relatively high values of 0.32 and 0.29 respectively.
That is to say, there is a considerably tendency for induced income to flow from the
backward areas to the advanced area.
The main reasons for this tendency are that:

(1) the sensitivity of dispersion

29

for industries shown in the Leontief inverse

is relatively high in the Middle, (2)

as shown by the coefficient values of matrix C in Table b of the Appendix, the


consumption coefficients of the Northeastern and Western regions for commodities
produced in the Middle are relatively higher than the Middle's consumption

coeffi~

cients for the Northeastern and Western commodities, and (3) as shown by the
coefficient values of matrix V presented in Table a of the Appendix, the ratios of
income-flow-out into the Middle from the Northeast and West are relatively higher
than those of the opposite directions.

The interaction among these coefficients in

the process of propagation account for the concentration of income in the advanced
area.

IV.

Campa~~on

06 Final Vemand and the Regional

Income-V~tkibutlon

The analyses in the preceding section have focused on the induced effects of a
unit of regional income, and have ignored the implications of the structure, or
commodity composition, of each region's autonomous demand.

In order to closely

examine this aspect, we have to shift our attention to the equation Y = KVBf.
We shall inquire into the problems of composition in two different ways.

In the

fundamental equation of income formation Y = KVBf, (i) let us replace the autonomous
demand vector f by the matrix Fl which shows autonomous expenditures separated into
three regional demands, and (ii) replace the vector f by the matrix F2 showing
autonomous expenditures separated into demand categories such as investment,
government expenditure and export, etc.

We thus get two types of direct and indirect

income-formation such that:


Yl = KVBFl

(2.8)

where, because the autonomous demand matrices Fl and F2 have the orders of nk x k
and nk x m respectively (m is the number of demand categories), the income Yl and Y2
become matrices having the orders of k x k and k x m respectively.
results are shown in the Table 2 and 3.

The estimated

30
The figures in Table 2-(a) are derived by dividing the column elements of Yl ,
which are shown in Column (c), by the corresponding regional total demands, and they
represent the coefficients of inducement to income per unit of autonomous demand by
each region.

The general features of Table 2-(a) are substantially the same as those

of Table 1, but between them there are some differences in detail.


is a marked contrast between the induced effects by

~eg~on

in the last line of the Table) and the induced effects by


in the last column of the Table).

As before, there

06 o~g~n (column totals

~eg~on 06 ~ec~pt

(averages

The results confirm our previous observation of

the concentration of income-formation in the Middle region.

However, compared with

the preceding Table, the degree of concentration, or the differential in coefficient


values of the Middle to those of the Northeast and the West, are magnified markedly
in this case (0.625 versus 0.117 and 0.190).

Clearly this result is produced by the

pattern of autonomous demand in each region, and reflects the fact that the location
of autonomous demand has a substantial effect in determining regional income generation, especially in the income-receiving base.
Table 2-(b) translates this result into the percentage dependency of income
formation by type of regional demand.

For the national averages indicated in the

last line of the Table, 68.3 percent of all the income comes directly and indirectly
from the initial expenditure in the Middle, and the contribution of expenditures in
the Northeast and in the West are only 11.4 and 20.3 percent respectively.

Further,

as shown in each of the row elements, their regional pattern exhibits a striking
contrast between the advanced and backward areas.

In the Middle area, a large

proportion of its income-generation depends on the demand originating in the Middle


itself (83.5%); the dependence on the Northeast and on the West are only 6.0 and
10.5 percent respectively.

The converse is true in the backward areas:

in the

Northeast and in the West the self-dependent ratio of income formation is only
50

60 percent, and a relatively large proportion of their income depends on

expenditure originating in the Middle, the advanced area of Japan.

The dependency

ratios on the Middle are 40.3 percent in the case of the Northeast and 35.4 percent
in the case of the West.
Table 3 is concerned with the direct and indirect income generation by final

31

TABLE 2
DIRECT AND INDIRECT INCDME FORMATION BY REGIONAL DEMAND *
(a) The coefficients of income-inducement per 1 unit of each region's demand **

demand
origin
of lncome
receipt

Northeast

Middle

West

Average

Northeast
Middle
West

0.597
0.368
0.072

0.066
0.733
0.095

0.045
0.354
0.623

0.117
0.625
0.190

Total

1.037

0.894

1.022

0.932

reglo~

(b) The percentage dependency of income-formation by type of regional demand **


Junit: rJ

demand
origin
of lncome
receipt

Northeast

Northeast
Middle
West
Average

reglo~

Middle

West

Total

52.6
6.0
3.9

40.3
83.5
35.4

7.1
10.5
60.7

100.0
100.0
100.0

11.4

68.3

20.3

100.0

(c) The direct and indirect income-formation due to regional demand


(unit: hundred millionyenj

demand
origin
of lncome
recei pt

reglo~

Northeast

Middle

West

Total

Northeast
Middle
West

6,457
3,976
775

4,942
54,942
7,099

870
6,897
12,157

12,270
65,815
20,031

Total

11 ,208

66,983

19,924

98,116

* Here, the regional demands are the tatals of final demands other than
endogenous consumption expenditure.
** Figures in (a) are derived by dividing the column elements of (c) by the
corresponding regional demands, and figures in (b) are calculated by dividing the
row elements of (c) by the corresponding tatals shown in the fourth column.

32

demand category.

Of course, the effects of endogenous consumption are included in

the multiplier side, while on the multiplicand side we include business consumption,
exogenous household consumption and government consumption as exogenous items.
Other exogenous elements are investment expenditure, i.e. gross domestic fixed
capital formation and net inventory change, and export demand.
total is the same as that of Table 2.

The exogenous demand

As shown by the figures in Table 3-(a), the

coefficients for the Middle area have excessively high values for all final demand
categories.
Looking at the last line of the Table 3-(a) and comparing Columns (1), (2) and
(3) with Columns (4), (5) and (6), the multiplier effects on the household incomes
for the national economy are somewhat higher in the case of exogenous consumption
items than in the case of investment items.

That the value of the consumption

multiplier is higher than that of the investment multiplier illustrates a


characteristic not verifiable under the Keynesian type of macro-multiplier model. 6 )
However, as shown in Table 3-(b), among the figures of percentage dependency by
demand category, the weight of investment demand is excessively high, reflecting its
relative importance in the composition of exogenous final demand expenditures.
In order to determine the degree of leakage from the backward areas to the
advance area by the type of demand category, we must have a breakdown of the induced
pattern in terms of the three regions.

This is provided in Table 3-(c).

Looking at

the columns for the Northeast and the West and comparing the relative ratios of
leakage to the Middle by each demand item, one finds that investment expenditure
(especially the fixed capital formation) has extremely higher leakage ratios
compared to those of the other autonomous demand items.

The self-sufficient ratios

of commodities needed directly and indirectly by investment are approximately 50


percent in the cases of the Northeast and the West.

By contrast, the corresponding

ratio of self-sufficiency of the Middle amounts to more than 80 percent.

This result

6) Except in the case of general government consumption, the multiplier effects of


each autonomous demand on incomes have values of less than one which may seem
strange. But this is solely because the income formation treated in our empirical
model is concerned with the household sector only and not with all value-added
sectors in the economy.

33
TABLE 3
DIRECT AND INDIRECT INCOME-FORMATION BY AUTONOMOUS DEMAND CATEGORY *
5

Bus i ness
Exogenous **
Gross
Net
Export Final
consumption household government domestic
inventory
demand
expenditure consumption consumption fixed capital change
total
ex enditure ex enditure formation
a The coefficients of income-inducement er 1 unit of demand
Northeast
0.135
0.152
0.178
0.100
0.120
0.080
Middle
0.633
0.586
0.691
0.609
2.555
0.651
West
0.191
0.254
0.272
0.169
0.173
0.153
Total
0.959
0.992
1.141
0.878
0.848
0.884
of income-formation b t
23.9
39.9
17.3
45.3
22.4
41 .2
19.2
43.6
(c) Di rect and i ndi rect income-formation by type of demand
unit: hundred million en
Northeast
1 ,071
2,931
4,889
1,372 12,270
814
Middle
4,108
11 ,388
29,765
3,760 11 ,191 65,815
West
1,783
4,488
8,259
1 174
2,635 20,031
Total
6,962
18,807
42,913
5,748 15,198 98,116
(unit: %
68.5
46.4
70.8
62.5
57.6
61.9
44.7
24.0
35.5
31.8
26.8
31.5
9.0
5.2
6.0
6.9
6.3
4.7
100.0
100.0 100.0 100.0
100.0
100.0
Northeast
Middle
West
Total

8.3
82.5
9.2
100.0

5.3
88.3
6.4
100.0

7.6
80.1
12.3
100.0

7.8
81.7
10.5
100.0

7.2
82.3
10.5
100.0

7.4
82.2
10.6
100.0

Northeast
Middle
West
Total

4.4
29.3
66.3
100.0

3.2
25.5
71.3
100.0

4.9
43.1
52.0
100.0

4.2
26.4
69.4
100.0

4.8
36.9
58.3
100.0

4.3
34.6
61.0
100.0

* The autonomous demand in this table excludes the consumption expenditure of


endogenous terms, and includes exogenous terms.
** The exogenous household consumption consists of the expenditure from the
transfer income from government, and its propensity to consume is assumed to be
equal to one.

34
TABLE 4
f1ATRIX MULTIPLIER OF INCOME-FORMATION (KVB)
1
Agriculture,
Forestry and
Fi sheri es

2
Coal and
Lignite
Mining

3
Mining
(except Coal
and Lignite
Mining)

4
Food and
Kindred
Products

5
Textile
Products

6
Lumber and
Wood
Products

I Northeast
Northeast
.906538
Middle
.243481
West
.053783
Tota
1.203802

1.009660
.280463
.054627
1.344749

.460262
.152632
.030566
.643460

.630688
.242529
.056604
.929821

.676045
.365511
.091897
1.133454

.843026
.268192
.054708
1.165926

Northeast
i~i ddl e
West
Total

Middle
.049332
.821486
.059352
.930169

.027020
.250411
.018038
.295469

.015286
.255333
.018153
.288772

.104287
.561904
.084368
.750559

.064992
.840818
.129003
1.034813

.081788
.821792
.111207
1.014786

III
Northeast
Middle
West
Total

West
.039351
.238634
1.004575
1.282560

.031125
.214965
.851031
1.097121

.008866
.061311
.236874
.307052

.041476
.228554
.689470
.959501

.048937
.445761
.652674
1.147373

.041067
.268799
.914037
1.223903

14
Metal
Products

15
Machinery
(n.e.c. )

16
Electric
Machinery

I Northeast
.532407
Northeast
Middle
.293503
West
".071810
Total
.897720

.415495
.274414
.056836
.746745

.358776
.402575
.061001
.822352

.478679
.380353
.079384
.938416

.386458
.313989
.048321
.748768

.665567
.344897
.083648
1.094

Northeast
Middle
West
Total

Middle
.057636
.691832
.089505
.838973

.052769
.584263
.075990
.713022

.057546
.630782
.076658
.764986

.054532
.683974
.089084
.827590

.054100
.700228
.068522
.822850

.087508
.736311
.102293
.926 12

III
Northeast
Middle
West
Total

West
.040977
.261317
.571499
.873393

.039003
.292234
.411371
.742608

.042746
.339861
.399158
.781866

.045068
.356402
.508611
.910081

.031320
.273683
.447222
.752226

.051917
.299476
.767488
1. 118881

II

II

18
19
17
Transporta- Precision Miscellaneous
tion
Machinery Manufacturing
Equipment

35

7
Pulp, Paper
and Paper
Products

8
Leather,
Leather Products and
Rubber
Products

9
Chemicals

10
Petroleum
and Coal
Products

11
Cerami c,
Clay and
Stone
Products

12
Primary
Iron and
Steel
Manufacturing

.710320
.267046
.057485
1.034851

.704887
.373968
.069501
1.148356

.542014
.236430
.057964
.836409

.488399
.166134
.043218
.697752

.655752
.254498
.056331
.966680

.440545
.191649
.037914
.670109

.427490
.181517
.036507
.645514

.112561
.718789
.136215
.967565

.062787
.783310
.088539
.934636

.060910
.551874
.096881
.709666

.032224
.228896
.047551
.308671

.062738
.688831
.089783
.841353

.059661
.432818
.078980
.571459

.059643
.361897
.063054
.484594

.050683
.266685
.678919
.996287

.041074
.334509
.718384
1.094566

.045674
.227787
.544705
.818166

.016850
.110659
.357799
.485308

.038872
.220726
.664984
.924582

.035173
.182297
.389012
.606481

.060441
.144228
.320252
.524921

20
Construction

23
21
22
Electricity, Wholesale Finance,
Gas, Water
and Retail Real Estate
and Sani tary Trade
and other
Services
Services

24
25
Transporta- Unallocated
tion
and Warehousing

.597559
.342581
.069564
1.009704

.577969
.175603
.034666
.788237

.957569
.250550
.047880
1.255999

.885472
.310845
.053373
1.249689

.813532
.260188
.066172
1.139892

.415535
.235111
.049605
.700251

.071192
.726509
.103628
.901329

.096956
.510829
.083456
.691241

.046641
.918089
.054238
1.018969

.055398
.978103
.065907
1.099407

.071111
.822386
.080381
.973878

.052801
.493789
.081708
.628298

.042013
.318394
.627697
.988104

.026387
.170733
.642278
.839398

.035126
.246350
1.005906
1.287382

.038302
.300031
.942282
1.280615

.069889
.254636
.735568
1.060093

.036963
.244505
.440008
.721476

13
Primary
Nonferrous
Metal
Manufacturi ng

36

suggests that a development program in the backward areas will have to contend with
many leakages in the propagation process and much of the benefits it generates in the
form of income will accrue to the advanced area.
It is instructive to examine how each sector's industrial activity plays a part
in the interregional income-formation process.

For this purpose, we must turn our

attention to the matrix multiplier KVB itself, which is shown in Table 4 having the
order of 3 rows and 75 (= 25 x 3) columns.
Inquiring into the first 3 x 25 part tabulated as the Northeast Column (I), we
can note which Northeastern industries have high leakage-ratios to the Middle area.
These are Electric machinery, Transportation equipment, Leather and rubber products,
and Textile products, all with values in excess of 0.35.

But instead of considering

the order of absolute values, if we take the relative ratios of leakage to selfinduced effects, Electric machinery stands out clearly from the other sectors, with
Precision machinery and Transportation equipment next in rank.
In the case of the West, shown as part (III) of the Table, the ranking of
sectors shows a somewhat different pattern compared with that of the Northeast.

The

highest leakage-ratio to the Middle is found in the Textile industry, and next in
Transportation equipment.

But for the relative leakage ratio, Electric machinery has

by far the highest value.

On the whole, the sectors having relatively low ratios of

self-induced effects are concentrated in some main manufacturing industries, and this
appears to be more a characteristic of the Northeast than that of the West.

v.

Output Ve.teJUn.i.nruon and InteJVLegiona1. Income Genvw..tion

The analyses that have been made so far were limited to the income formation
process with the output determination side appearing only in an indirect manner
through the operation of the income multiplier as a matrix.

We shall now directly

study the output determination side and compare briefly the analytical results based
on two different models:

namely, the model exo that regards the consumption demand

as an exogenous variable and the model endo that regards it as endogenous.

These

37
analytical systems have already been given in equations (2.7).7)
Table 5 is a summary version of this comparison.

The figures in columns "exo"

were computed by using the original Leontief inverse B, and "endo" were calculated by
using the enlarged inverse B[I - CVB]-l or B[I + CKVBJ.

It should by noted that the

final demand as a multiplicand differs in the two cases:

in the case of the former

it is f

f, and in the case of the latter it is f, where, of course, f a and fare

vectors of the endogenous consumption demand and the final demand excluding
endogenous consumption respectively.
With regards to the effects of final demand by the demand category indicated in
Column I in the Table, the direct and indirect requirements of output in the
endogenous model have higher values (except the household consumption column (2))
than in the case of the exogenous model.

But, by comparing these two cases more

closely it is interesting to see what particular types of final demand have more
influence through endogenous changes in consumption demand.

This can be found by

dividing the figures of the endo-columns by the corresponding figures of the


exo-columns.

In addition, if we breakdown the effects in terms of industrial

sectors, we may observe which sectors are relatively more affected by consumption
demand.

It may be observed that there are differential effects in each sector by

regions also.
Next, let us compare the effects of regional demands in the two models.
are indicated in Column II of the Table.

These

As shown in the diagonal blocks of Column

II-(a), it is noteworthy that the effects on direct and indirect output requirements
to the OWn

~egion

clearly decrease in the case of the endogenous model as compared

with those of the exogenous model.

The same is true for the percentage dependency

of output requirements indicated by the diagonal blocks of Column II-(b).

In other

words, the transformation from an exogenous to an endogenous model reduces a region's

7) The first applications of interregional input-output analysis that took into


account the endogenous income sector were given by H. B. Chenery's Italian
two-region model and L. N. Moses' American three-region model. See, H. B. Chenery
[5J, L. N. Moses [38J. However, in neither model, was the role of income
formation separated out from the output-determination mechanism in an explicit
formal framework.

38

TABLE 5
COMPARISON OF DIRECT AND INDIRECT REQUIREMENTS OF
OUTPUT BETWEEN THE TWO MODELS *
Final demand by demand category

1.

(1)

(2)

Household
consumption**
expenditure
total exogenous
only

Business
consumption
expenditure

exo

el1do

exo

evtdo

(3)

General
government
consumption
expenditure

exo

evtdo

(4)

Gross domestic
fixed capital
formation

exo

(a) The direct and indirect requirements of outputs

evtdo

Northeast
Middle
West

1 ,936
11,914
2,986

3,737 20,982
20,420 106,860
5,665 34,521

3,181
14,448
5,624

3,640
17,844
6,123

7,871
35,910
12,817

Total

16,836

29,822 162,363

23,252

27,607

56,598 125,410 190,745

10,456 18,487
91,527 135,559
23,427 36,699

(b) The percentage dependency of output requirements


Northeast
Middl e
West

4.7
4.4
3.9

9.1
7.5
7.3

51.1
39.5
44.7

7.8
5.3
7.3

8.9
6.6
7.4

19.2
13.3
16.6

25.5
33.8
30.3

45.1
50.1
47.5

Total

4.3

7.7

41. 7

6.0

7.1

14.5

32.2

49.0

(c) The coefficients of output requirements per unit of demand


Northeast
Middle
West

0.219
1.346
0.337

0.422
2.307
0.640

0.248
1.263
0.408

0.453
2.061
0.802

0.221
1.082
0.371

0.477
2.178
0.777

0.214
1.871
0.479

0.378
2.772
0.750

Total

1.902

3.369

1. 919

3.316

1.674

3.432

2.564

3.900

* The model exo regards the consumption demand as exogenous, and the model
el1do as endogenous; their solutions are given by X = [I - A]-l{fc + f} and

[I - A]-l[I - CVB]-l f

B[I + CKVB]f respectively.

** The consumption in Column-"total" in (2) includes the term f c ' and that in
Column-"exogenous onl y" excludes f c .

39

II.
(5)

Net
inventory
change

exo

I ettdo

(6 )

Final
demand
total

Export

exo

I ettdo

exo

I ettdo

(i)
Northeast

exo

I ettdo

Final demand by region


(i i i )
West

(i i )
Middle

exo

J ettdo

exo

J ettdo

(unit: hundred million yen)


1,577 2,900 2,434 4,950 41,026 41,026 26,403 20,759 12,741 17,229 1,881 3,038
10,921 16,655 31,741 47,815 270,807 270,807 14,847 16,987 232,375 224,823 23,585 28,997
2,856 4,699 7,319 11,728 77 ,232 77 ,232 2,509 3,328 23,379 30,722 51,34443,182
15,354 24,154 41,495 64,493 389,065 389,065 43,760 41,074 268,495 272,77576,81075,217

%)

19.3

3.8
4.0
3.7

6.8
6.2
6.1

5.9
11. 7
9.5

12.1
17.7
15.2

100.0
100.0
100.0

100.0
100.0
100.0

64.4
5.5
3.2

50.6
6.3
4.3

31.1
85.8
30.3

(unit:
4.6
42.0
83.0
8.7
39.8 66.5

3.9

6.2

10.7

16.6

100.0

100.0

11.2

10.6

69.0

70.1

0.233 0.413 0.142 0.288


1.612 2.458 1.847 2.782
0.421 0.694 0.426 0.683

0.224
1.481
0.422

0.391 1.230 1.920


2.571 0.691 1.574
0.734 0.117 0.308

0.102
1.863
0.187

0.230 0.051 0.156


2.999 0.644 1.486
0.410 1.404 2.214

2.266 3.565 2.415 3.753

2.128

3.696

2.152

3.659 2.099 3.856

2.038 3.798

19.7

7.4
10.7
55.9

40

own inducement powers and enhances those of the other regions.


drawn from this fact is as follows:
be liable to

ov~vaiue

A conclusion to be

by using the customary exogenous model, we may

the effects on the own region's output requirements.

With regards to the effects of regional demands on the national economy, the
results are shown in the last line of Column II-(a) or -(b).

For the endogenous

model, in contrast to the exogenous model, the direct and indirect output requirements due to demand originating in a backward region (Northeast or West) is lower in
value, while the value due to the demand originating in the advanced region (Middle)
is higher.
things.

Nevertheless, the powers of inducement itself invert this order of

As shown in the last line of Column II-(c), in the case of the endogenous

model, the coefficient of output requirements of the Middle takes the lowest ranking
in order.

Just the opposite is true in the case of the exogenous model in which the

coefficient of the Middle is shown to have the highest value among the three regions.
In short, switching the model from exogenous to endogenous means that the output
requirements due to the autonomous final demand originating in the advanced area
increase in ah60lute value, but that the

deg~ee

of increase itself is not as large as

the degree of increase in each of the backward areas.

These results are attributable

to the working of the "subjoined inverse multiplier" [I + CKVB] in the propagation


process of output determination.

41
APPENDIX
TABLE a
COEFFICIENTS OF VALUE-ADDED OF THE HOUSEHOLD SECTOR IN EACH REGION
(unitl0- 4 )
Agricul Min- Light Heavy Con- Trade Finance Trans- Others Average
Serv- portai nstructure,Fol" ing i ndustry dustry tion
ices tion
estry
Northeast
Northeast
Middle
West
Total
(Output)

4,638 4,871
8 181
0
7

1,009 1,983 4,967 4,884 4,216 1,059


36
50
69
36
81
9
2
0
0
0
0
0

2,983
43
1

4,646 5,059 1,044

1,061

3,027

8,277 1,436 8,392

6,400 3,594 2,401

996
44
3

2,052 5,003 4,920 4,297 1 ,068


6,941

1,721

1,863 41 ,026

Middle
Northeast
Middle
West

12
5,386 4,470 1,342
1
1
2

3
0
0
1,377 1,854 4,555 4,652 3,415 8,388
3
2
2
0

2,412

Total

5,388 4,484 1,344

1 ,379 1,856 4,561

2,414

(Output)

14,503

4,654 3,418 8,388

917 62,163 88,105 22,314 18,93840,694 11,593 11,580 270,807


West

Northeast
Middle
West

0
0
108
21
23
5,296 4,440 1,140

Total

5,317 4,549 1,163

(Output)

2
0
2
4
86
86
95
5,076 4,934 3,520 1,057

42
2,588

1,020 1,967 5,163 5,022 3,617 1,062

2,631

25
47
994 1 ,919

9,513 1,582 15,075 23,360 5,906 3,553 11,986 2,824 3,432 77 ,232
National average

Northeast
Middle
West
Total
(Output)

98
983
202

450 1,169
55
225
482
569
1,036 1,317 3,481 3,196 2,479 5,776
198
726
993
618 2,151
357

315
1,692
515

5,177 4,720 1,283

1,289 1,899 4,689 4,759 3,547 9,096

2,522

1,189 1,781
2,427 1,151
1,561 1,788

32,294 3,934 85,631 117,865 31,815 24,892 59,620 16,138 16,874 389,065

* Sectors listed here are aggregated to 9 from the original 25 classifications


for convenience.
** Unit of output: one hundred million yen.

42
TABLE b
CONSUMPTION COEFFICIENTS OF EACH REGION
Northeast
I~ortheast

Agri cu lture & Forestry


Mining
Light Industry
Heavy Industry
Constructi on
Trade
Finance &Services
Transportation
Others
Sub-total
Middle
Agriculture & Forestry
r,1ining
Light Industry
Heavy Industry
Construction
Trade
Finance & Services
Transportation
Others
Sub-total
West
Agriculture & Forestry
Mining
Light Industry
Heavy Industry
Construction
Trade
Finance ~ Services
Transportation
Others
Sub-total
Total
Agriculture & Forestry
Mining
Light Industry
Heavy I ndus try
Construction
Trade
Finance & Services
Transportation
Others
Total

Middle

(unit:10- 4 )
Nati onal
I average

West

764
76
2,830
74
0
1,255
2,151
409
140
7,699

87
1
156
5
0
5
8
1
5
268

42
0
43
2
0
2
1
0
0
90

170
11
497
13
0
175
300
57
22
1,245

113
0
1,305
499
0
157
46
11
2
2,133

429
3
3,789
542
0
1 ,321
2,496
600
259
9,439

47
0
961
423
0
165
55
4
0
1,655

302
2
2,828
510
0
907
1,623
389
168
6,729

24
0
133
6
0
4
1
0
0
168

95
2
170
12
0
4
8
2
0
293

612
5
3,388
108
0
1 ,103
2,398
503
137
8,255

199
3
873
32
0
245
533
111
30
2,026

901
76
4,268
579
0
1 ,416
2,198
420
142
10,000
(0.870)

611
6
4,115
559
0
1,330
2,512
603
264
10,000
(0.761)

702
5
4,392
533
0
11 ,270
2,454
507
137
10,000
(0.855)

671
16
4,198
555
0
1,328
2,455
557
220
10,000
(0.794)

* Figures in parentheses in the last line show the total propensity to consume
of each regi on.
** Sectors listed here are aggregated to 9 from the original 25 classifications
for convenience.

CHAPTER 3
FOREIGN TRADE MULTIPLIER, INPUT-OUTPUT ANALYSIS
AND CONSUMPTION FUNCTION*

I.

Inttoducton

There has long been certain defects in the treatment of imported intermediate
goods in the usual Keynesian foreign trade multiplier analysis.

In its fundamental

equation for an open economy, Y + M = C + I + E, the external sector is combined


inconsistently with the domestic sector in the circular flow.

Y stands for net

national product excluding intermediate goods, while M stands for imported goods.
including intermediate products.
On the other hand, Leontief's matrix multiplier model has been devoted exclusively to the analysis of intermediate products in the circular flow; but unlike the
Keynesian model, it lacks a mechanism for the multiplier process via the consumption
function.

Formally, the household sector can be dealt with in the Leontief system

as an industry whose output is labor and whose inputs are consumption goods.
However, a more appropriate procedure in dealing with consumption is not to regard it
as a fictional production activity but to introduce the Keynesian consumption
function in its disaggregated form.
In this chapter, we shall try to close these two gaps by revising the conventional foreign trade multiplier, on the one hand, and by developing an input-output
analysis dealing with the consumption function, on the other. Thus, without
considering the income-distribution structure, we extend our input-output model of
Chapter 1 to an open economy.
It was shown in Chapter 1 that in the case of a closed economy with no foreign
trade, the conclusion that total income varies in accordance with the proportions of
final demand can be obtained only by considering the structure of income distribution
together with the input-output structure.

In the case of an open economy with

* This chapter appeared originally, under the same title, in

The Qu~~y JOU4nal


06 Eeonomic6, Vol. 74, No.1, Feb. 1960, and is reproduced with minor revisions
and with some additions to the Tables. The author is indebted to Mr. Shingo
Masegi, professor of Yokohama City University, for the mathematical formulation in
Sectlon VI.

44

foreign trade, however, the same conclusion can be derived without explicitly
considering the distribution structure.

This is due to the fact that the coeffi-

cients on the production side enter into the income formation process through
imports.

And this is exactly that we shall develop in this chapter.

In the traditional Keynesian foreign trade multiplier, the income-generating


process is usually given as:
2

+ P + P + P + ..

1
=~

(3.1)

where p is the marginal propensity to consume domestic goods.


preceding each component of the series is a
tion.

But underlying and

~ub-muttlpteA ~oc~~

involving produc-

For example, the first item of the above series (the primary additional income

1), is the outcome of the production activity in the sectors favored with the primary

round of one unit of autonomous injection (e.g., autonomous investment or export).


This primary subprocess due to one unit of autonomous injection can be expressed as
follows:
Increase in Y (= income)
2

(1 - a) + a(l - a) + a (1 - a) +

Increase in R

(=

a + a 2 + a3

1 - a

=.,.---:-a=

(a)

intermediate goods)
=_a_

(b)

1 - a

where (1 - a) is the value-added ratio, i.e., a is the total input coefficient.

But

this procedure is appropriate only in the case where there is no leakage in the
submultiplier process, and in orthodox Keynesian reasoning this condition is satisfied by the assumption that the multiplicand is equal to the autonomous investment
(or export) minuh import of intermediate goods required for the production of
investment goods (or exported goods).

Hence in the usual Keynesian analysis, the

imported intermediate goods required for the production of investment goods (or
exported goods) are treated as an exogenous factor in the multiplier process.
Logically, however, we should treat the imported intermediate goods as an endogenous
factor induced by the initial injection. Thus we should rewrite the above

45

submultiplier process as follows:


Increase in Y
2 2

(1 - a) + ay(l - a) + a y (1 - a) + ....

1 - a
= -1-- ay

(a)'

Increase in R
2
3 2
a+ay+ay +

= __a_

(b)'

1 - ay

where y is the self-sufficing ratio of materials (that is, y = B/R) , where B is the
demand for demestic intermediate products.

~l
- a
-ay

<

Obviously, since y

<

1, we have

1, i.e., 1 unit of autonomous injection does not create an equal amount of

-- ay
a = h, then we can call 1 - h
primary additional income, but less than 1. Let ~l
the leakage coefficient in the submultiplier process dealing with production
activities.
Since similar submultiplier processes precede all the other secondary increases
in income (due to additional consumption expenditures), the whole income-generating
process can also be given as:
(3.2)

This foreign trade multiplier takes into account the intermediate products in the
circular flow.
Of course, the usual Keynesian foreign trade multiplier generally does take
into account the import of intermediate goods required for the production of
consumption goods, but this is done inadequately.

For, in the standard Keynesian

analysis, the entire propagation process is projected into the income-spending or


expenditure side only, and no

~xplieit

distinction is made between the leakage in

the expenditure process and the leakage in the production process.

Nevertheless,

the intermediate goods required for the production of consumption goods, as well as
those required for the production of investment (or export) goods, are not imported
at the income-expenditure level, but in the production-submultiplier process.

In

our multiplier, the import of intermediate goods is taken into account in its proper
context, namely, in the circular flow of intermediate products.
In order to express our multiplier in a form comparable with the orthodox

46

Keynesian multiplier, let X = Y + R denote the total output, (1 - a) = Y/X the value
added ratio, and A = R/Y the material coefficient, then A = ~
= R/X = ~-so that
Y Y/X
1 - a'
1 - a

- a

h=-l_ay=l - a + a -

+ All _

ay

y). substituting the above relation in (3.2),

our foreign trade multiplier becomes


__h_ =
1
or 1 - ph
1 - p + A( 1 - y)
- -:-1--~(a---m'"T")-+~A..-.(1---y-'-)

(3.3)

where a stands for the marginal propensity to consume and m the marginal propensity
to import finished goods (p = a - m).

III.

The Modi6ied

Multip!i~

and

~he

Fundamental Equation

6o~

an Open Economy

From this new foreign trade multiplier (3.3), we derive the following conclusions:
(i)

Depending on the industries that enter into the propagation process, our

multiplier has different values since the interindustry average values of A and y
differ with each pattern of propagation. This is a characteristic which is not
found in the orthodox Keynesian foreign trade multiplier.
(i i)

Let us set y = 1 in (3.3); our multiplier is then equal to -1_1- or


- p

1 _ (a1 _ m) , and coincides with the Keynesian foreign trade multiplier in the case
where induced imports are restricted to finished goods only.

In this case, the

Keynesian marginal propensity to consume domestic goods, i.e., p in series (3.1),


coincides with our concept of p in series (3.2), since with no imported intermediate
goods, h becomes equal to 1 and series (3.2) is converted into series (3.1).
(iii)

If the induced import in the Keynesian model also includes intermediate

goods required for the production of consumption goods, the Keynesian concept of p
will differ from our concept.

The Keynesian p in (3.1) espresses the relation

between income and that consumption imputable to domestic factors, in other words;
the value of imported intermediate goods is excluded from the value of consumption.
But this is clearly a nonoperational concept.

Such a nonoperational concept is

necessary in the Keynesian model because all propagation processes are projected into
the income-expenditure side only.

In our multiplier, the process is decomposed into

47
two distinct propagation processes, assuming essentially different characteristics.
Our multiplier also can be derived from a revised fundamental equation for an
open economy.

We rewrite the Keynesian equation as follows:


Y+M

=C+I+E

Y+tI+~

=C+I+E

r--"--\

Y + (R - H) + MC = C + I + E
where
R

tI = the

= total

(3.4)

imports of intermediate goods, ~ = the imports of finished goods,

intermediate goods required for the production of net national product,

H = the demand for domestic intermediate products.

Equation (3.4) is our fundamental

equation for an open economy.


Here we can incorporate the import function of intermediate goods by using two
parameters, the material-coefficient A. and the self-sufficing ratio of materials y,
that is

tI = (R - H) = (1

(3.5)

- y)AY

The economic meaning of this function is that the import of intermediate goods is
not directly a function of Y, but rather a function of AY, i.e., the total intermediate goods required for the production of yl).

The import function of finished

goods and the consumption function may be written as ~ = ~(y) and C = c(y)
respectively.

By combining our fundamental equation (3.4) and these three functions,

we get
Y + A(l - y)Y + ~(y) - c(y) = I + E

(3.6)

Differentiating the above equation, we obtain


6Y/6(I + E)

= 1 _ (0 - m) + A(l - y)

-p+A(l-y)

(3.7)

This result coincides with (3.3) derived by working out the geometrical progression which combines the income-circular flow with the circular flow of intermediate
1) The import function of intermediate goods may also be written as:
tI = (1 - y)AY + g
where g is constant. In this case, parameters y and A become "marginal"
coefficient.

48

products.
The parameters y, A, a, which are crucial in our foreign trade multiplier
formula, may have different values depending on the nature of the industries entering
into the propagation process. Therefore, for a more detailed analysis, we must
utilize Leontief's analysis of interindustry relations.
For a comparison with the Leontief system, let ~ = ~/X denote the import
coefficient, then A(l - y) in (3.7) may be converted into the following expression:
A(l - y)

= ~Y = ~/X
= --1:!....-..
Y/X
1 - a

Substituting this relation in (3.7), we obtain another variant for our foreign trade
multiplier:
tJ.Y/tJ.(I + E)

1 - a
(1 - p)(l - a) + ~

(3.8)

Now, if we take the total output multiplier instead of the income multiplier (3.8),
we get
1

AX/tJ.(I + E) = (1 _ p)(l - a) + ~

(3.9)

This total output multiplier (3.9) corresponds to a Leontief inverse matrix in which
the consumption function is integrated. 2) In order to prove this point, we must
turn to the input-output system in the next section.

IV.

IntvUndwdJLY Ana.y.6-U. and .the Con.oump:Uon Functicm:

A Ma.tJL.i.x MuU-tplieJt

Comb,[n..i.ng Leont,[e6'.6 PItOpa.ga;Uon PItOc.e.o.6 and .the Keyne.o,[an PItOpa.ga;Uon

We write the input-output system as follows:


[I - A + M]X = f a + f

(3.10)

2) Since we define X as Y + R, X may be interpreted as domestic total output plus

imports. But if a system expressed in terms of domestic output is preferred,


X may then be redefined as equal to Y + domestic intermediate goods, with
corresponding modifications of coefficients a and ~. using the redefined X as
denominators. Similar procedures may be applied to the input-output system in
the next section.

49

where
I

= the unit matrix of

= the square matrix of input coefficients a1-J


..

n order

M = the diagonal matrix of import coefficients

~i

x = a column vector of outputs


f a = a column vector of consumption demand
f = a column vector of final demand other than consumption.
This model differs from that of Leontief only in the treatment of imports as an
autonomous element.

Now, if we regard fa as an exogenoU6 variable, as well as f,

the solution of (3.10) will be

X = [I - A + M]-lf
where

f = fa

(3.11 )

f. The marco-multiplier corresponding to this Leontief matrix

multiplier [I - A + M]-l in (3.11), can be derived by adding the two subseries {al'
and (b)' in the submultiplier process on page 45, i.e.,

..l...:....sf +
1-

ay

_a_ = _ _
1_ =

1-

ay

1 - ay

-=----=-__3)
- a +

But if we treat consumption demand fa as an endogenoU6 variable, we cannot


regard the household sector as autonomous and independent of the other n processing
sectors.

In such a case, following the usual procedure, we transfer the household

sector to the processing sectors, and the (n x n) matrix multiplier in (3.11) is


converted into a (n + 1) x (n + 1) matrix multiplier. The household sector is
accordingly regarded as an industry whose output is labor and whose inputs are
consumption goods, and is treated just as the others are, i.e., its input-output
ratios are assumed constant.

But consumers are not a technologically determined

production process, but are rather choice-making organisms.

Furthermore, the factors

of choice-making, i.e., the consumption coefficients, are not as stable as the input
coefficients of the other processing sectors, so that (n + 1) x (n + 1) inverse

3) This is because

ay =

RH
H
R-P1
xIi
= X = -X-- = a

- ~

50

matrix multiplier in the above procedure may embody an element of instability.


Therefore, instead of treating households as an industry, we should introduce
the Keynesian consumption function in its disaggregated form and retain the same
number of processing sectors as before.

Now the consumption demand f a can be

written as follows:

= C(I

fa

(3.12)

- A)X

where C =

the square matrix of the consumption coefficients.


n .... an

It is easily verified that C(I - A) equals the product of the column vector of
consumption coefficients and the row vector of value-added ratios. 4 ) Instead of
(3.12), if we add the nonhomogeneous term to the consumption function, a. becomes a
~

marginal coefficient.
included in

In this case, we can treat the nonhomogeneous term as

f.

Substituting the above consumption function (3.12) in (3.10), we get

[(I - C)(I - A)

M]X = f

(3.13)

Solving (3.13) for X, we obtain

X = [(I - C)(I - A)

M]-lf

(3.14)

This inverse matrix in (3.14) obviously corresponds to the one-sector multiplier


(3.9).

Further, if we write ~ - A)(I - A + M)-l


trix in (3.14) is converted into the form [I -

= H, we can show that the inverse maA +

M]-l[I - CH]-l -- i.e., the Leontief

inverse [I - A + M]-l multiplied by the inverse [I - CH]-l which shows the effects
4) Denoting by a the column vector of consumption coefficients and by i' the row
vector = (1, 1, ... 1), we then have C(I - A) = ai' (I - A). Since i' (I - A) =
(1 - Ia'l'
1 - Ia.~ 2 , .... , 1 - Ia.w ), the term i'(I - A) equals the row vector of
~
value-added ratios.

51
of endogenous changes in the consumption demand of the household sector. 5 ) The
proof is as follows:

let us write [I - A + M]-l = B, then the inverse matrix in

(3.14) turns into


[(I - C)(I - A) +

Mr 1

[I - A

M - C(I - A)]

-1

-1-1 = B[I - CH] -1


= [{I - C(I - A)B}B]

(3.15)

We can refer to the inverse [I - CH]-l as the "subjoined inverse matrix showing the
effect of endogenous changes in consumption. ,,6)
Since the above procedure which distinguishes the inverse

(reflecting produc-

tion activities) from the inverse [I - CH]-l (reflecting consumption activities)


obviously satisfies theoretical requirements, we can now combine these two inverses
by the formula (3.15).

This formula (3.15) can be interpreted as the multiplier

combining Leontief's propagation process and the Keynesian propagation process.

Numerical estimates for the above inverse matrices are presented in this
section.
60~

These estimates are based on data from TabtC6

06

Int~ndUh~

RetatiOn6

1951 and 1954, published by the Japanese Ministry of International Trade and

Industry,7) which contains an inverse matrix table corresponding to our inverse

5) We may designate the matrix H as the value-added m~pti~ by regarding the


consumption demand as exogenous, since i'H equals the product of the row vector
of value-added ratios and the Leontief matrix multiplier.
6) Precisely speaking, in the above model, C is the square matrix of consumption
coefficients of domC6t1c goodo, and M is the diagonal matrix of import coefficients of int~ediate goodo. But, if the distinction between imported
intermediate goods and imported finished goods is not made statistically, we must
then interpret C as the matrix of the consumption coefficient including imported
finished goods, and M as the matrix of the import coefficient of intermediate and
finished goods. Of course, this is not theoretical.
7) Ministry of International Trade and Industry (MITI)[23], which is fue first publication of input-output tables in Japan. In this volume, the concept of import
coefficient is defined to include imported finished goods as well as intermediate
goods. See preceding footnote 6).

52
TABLE 1
TOTAL DIRECT AND INDIRECT IMPORT REQUIREMENTS IN JAPAN
(1 )

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

Agriculture ( Food)
Agriculture (Textile Material)
Agri cul ture (Rubber t1ateri a1 )
Agriculture (Hide)
Forestry
Fi shery
Coal and Lignite
Crude Petroleum and Natural Gas
Iron Ores
Nonferrous Meta11 ic Ores
lionmeta 1i cOres
Food Processing
Textile and Apparel
Lumber and Timber Products
Pulp, Paper, Print and Publishing
Chemicals
Coal Products
Petroleum Products
Rubber Products
Leather and Leather Products
Stone, Clay, and Glass Products
Iron and Steel Products
Nonferrous Metals
t1achi nery
All Other Manufacturing
Manufactured Gas
Electric Power
Transportation and Communication
Construction
Trade
Business and Personal Service
Textile Rug and Waste
Scrap (Iron and Steel)
Scrap (Nonferrous Meta 1)
Scrap (Others)
Una 11 ocated

1951

Rate of
1954 Increase 1951

0.183
0.857
1.000
0.816
0.052
0.097
0.125
0.852
0.870
0.298
0.404
0.144
0.372
0.049
0.119
0.156
0.111
0.604
0.478
0.525
0.106
0.127
0.161
0.127
0.119
0.093
0.071
0.089
0.073
0.033
0.057
0.427
0.038
0.057
0.077
0.150

0.215
0.873
1.000
0.801
0.055
0.114
0.200
0.935
0.882
0.301
0.435
0,159
0.324
0.058
0.128
0.170
0.172
0.669
0.475
0.538
0.152
0.185
0.238
0.194
0.158
0.163
0.096
0.103
0.101
0.045
0.071
0.193
0.166
0.247
0.079
0.205

(%)

117.5
101.9
100.0
98.2
105.8
117.5
160.0
109.7
101. 4
101.0
107.7
110.4
87.1
118.4
107.6
109.0
155.0
110.8
99.4
102.5
143.4
145.7
147.8
152.8
132.8
175.3
135.2
115.7
138.4
136.4
124.6
45.2
436.8
433.3
102.6
136.7

0.336
0.883
1.000
0.850
0.229
0.266
0.288
0.879
0.894
0.429
0.515
0.304
0.486
0.227
0.279
0.314
0.278
0.678
0.575
0.614
0.273
0.290
0.317
0.290
0.283
0.264
0.246
0.260
0.244
0.214
0.231
0.532
0.215
0.231
0.247
0.307

(II)

Rate of
1954 Inc(~)se
0.405
0.904
1.000
0.849
0.285
0.329
0.394
0.951
0.911
0.471
0.572
0.364
0.489
0.287
0.238
0.372
0.373
0.750
0.603
0.651
0.358
0.382
0.423
0.390
0.362
0.369
0.316
0.321
0.320
0.277
0.297
0.393
0.369
0.431
0.303
0.428

120.7
102.4
100.0
99.9
124.6
124.0
136.7
108.2
101.8
109.8
111.0
119.7
100.6
126.6
85.2
118.4
134.5
110.6
104.9
106.0
131.2
131.9
133.3
134.5
127.8
139.8
128.2
123.7
131.1
129.9
128.9
73.8
171.8
186.6
122.9
139.7

53

TABLE 2
TOTAL DIRECT AND INDIRECT EMPLOYMENT (man)
REQUIREMENTS FOR THE OUTPUT OF ONE MILLION YEN
I!lbl

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

1954
23,468
16,925

111l7TIT
1.51
1.08

13,620
2,990
4,740
3,383
2,260
2,176
2,978
2,950
4,480
2,720
4,831
3,240
3,690
3,012
539
2,210
3,590
3,700
2,850
2,095
3,670
4,450
2,906
2,379
3,203
5,752
4,782
4,250
2,450
2,920
1,206
1,640
3,450

15,639
12,583
13,720
11 ,472
2,922
3,370
10,073
8,662
12,952
9,581
14,393
7,809
12,105
11,395
3,716
7 ,563
8,281
12,286
11,080
9,814
11 ,837
12,966
11 ,425
11 ,529
12,313
14,813
14,448
13,672
10,644
11 ,383
5,970
10,990
11,480

1.15
4.21
2.89
3.39
1.29
1.55
3.38
2.94
2.89
3.52
2.98
2.41
3.28
3.78
6.89
3.42
2.31
3.32
3.89
4.68
3.23
2.91
3.93
4.85
3.84
2.58
3.02
3.22
4.34
3.90
4.95
6.70
3.33

(II )

(II)/(I)

(I)

16,849
16,853

25,827
18,339

1.53
1.09

15,550
15,640

17 ,553
2,939
4,847
3,850
2,719
3,133
4,242
4,306
5,539
3,761
5,488
3,788
4,696
3,653
854
2,961
4,139
4,514
3,654
2,544
4,409
5,346
3,556
3,380
4,512
6,439
5,571
5,556
2,855
2,670
1,414
1,905
3,589

19,569
13,330
14,772
12,906
4,221
4,554
11 ,947
10,848
14,946
7,415
15,956
13,481
13,991
13,445
5,197
8,791
9,364
14,328
13,236
11,740
13,987
15,017
13,611
13,689
14,657
16,513
16,191
15,909
9,143
13,222
11 ,795
12,034
12,928

loll
4.54
3.05
3.35
1.55
1.45
2.82
2.52
2.70
1.97
2.91
3.56
2.98
3.68
6.09
2.97
2.26
3.17
3.62
4.61
3.17
2.81
3.83
4.05
3.23
2.56
2.91
2.86
3.20
4.95
8.34
6.32
3.60

(I)

(II)

54
TABLE 3
TOTAL DIRECT AND INDIRECT CAPITAL
REQUIREMENTS FOR THE OUTPUT OF ONE MILLION YEN

2
3
4
5
6
7
8
9
10
11

12
13

14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

(I )
1.8892
.3287
.4241
.6466
.9525
1.0828
.1836
.1455
.7710
.7865
.9631
.9132
.9330
1.0942
1.3756
1.6443
.3845
.7087
.6450
1.2204
1.3554
.864?
1. 1799
1.1092
2.1209
4.9968
3.5677
1.1957
1.2205
4.1621
.3643
.5713
.2946
.3893
1.3050

1951

4.4275
.7602

PI J/lI )
2.34
2.31

1 .8163
.2919

1954
{II J
4.8461
.7824

.9939
3.5860
3.7574
3.7949
.5383
.5473
2.9486
2.6355
3.6216
2.7997
3.8916
3.8336
4.0026
4.4117
1.6128
2.3282
2.1853
3.9939
4.0634
3.4643
3.8868
3.8424
4.9626
7.9102
6.4066
4.0428
4.2218
7.0881
2.1414
3.5536
3.2286
3.2519
3.9445

2.34
5.55
3.94
3.50
2.93
3.76
3.82
3.35
3.76
3.07
4.17
3.50
2.91
2.68
4.19
3.29
3.39
3.27
2.99
4.01
3.29
3.46
2.33
1.58
1.80
3.38
3.46
1. 70
5.88
6.22
10.96
8.35
3.02

.4580
.6872
1.0089
1.0686
.0833
.1368
.8027
.8947
.9384
.7890
1. 1492
1 .2107
1.3616
1.7201
.3406
.7044
.6728
1 .2747
1.3634
.9335
1.1598
1 .3381
2.2182
4.9597
3.5299
1.2378
1.3020
4.2831
.5515
.5320
.2516
.4163
1. 7197

1.2296
3.9549
4.4377
4.1571
.3361
.5927
3.5119
3.0756
4.1963
3.4355
4.8004
2.9552
4.5748
4.9302
1 .6299
2.7482
2.4640
4.5531
4.5059
3.8813
4.2783
4.5900
5.4956
8.4553
6.9034
4.6420
4.8920
7.7757
3.5897
3.6692
3.1813
4.0660
4.7941

(II )

(I)

(Million yen)
{II)/{IJ
2.67
2.68
2.68
5.76
4.40
3.89
4.03
4.33
4.38
3.44
4.47
4.35
4.18
2.44
3.36
2.87
4.79
3.90
3.66
3.57
3.30
4.16
3.69
3.43
2.48
1.70
1.96
3.75
3.76
1.82
6.51
6.90
12.64
9.77
4.36

55
In Table 1, we present the total direct and indirect requirements of imports
due to an increase in demand. The calculation was done in two ways.

Column(I) in

the table shows the import requirements computed by the use of M and the inverse
[I - A + M]-l in (3.11); these figures do not take into account the effects of

endogenous changes in consumption. On the other hand, the figures in column(II)


show the import requirements for which the effects of endogenous changes in consumption are considered. These figures were calculated by the use of M and the inverse
[I - A + M]-l [I - CH]-l in (3.15).
differs in the two cases:
column(II) it is

It should be noted that the final demand

in the case of column(I) it is

f. Comparing columns(I) and

f,

and in the case of

(II), we may determine which sector's

imports are relatively more affected by endogenous consumption demand. Similar empirical calculations for total direct and indirect requirements of employment and of
capital (Tables 2 and 3 respectively) show that there are some significant differences in each sector's factor response to endogenous changes in consumption demand.

VI.

Fo.l!JTll.li.a. 60Jr. :the Compu;ta;t<.on 06 :the Subjoined InveMe Showing :the E66ect 06
EndogenoU6 Change.h in COn6ump:Uon

If the consumption coefficients are not as stable as the input coefficients, it


is desirable that the "subjoined inverse" be converted into a form which is easily
computable and revisable.

In this section we shall develop a practical computation

formula for this subjoined inverse without inversing the matrix.

By our formula we

can renew the subjoined inverse at any time when it is necessary. This convenient
computation formula can also be derived more convincingly by tracing the propagation
process from the initial injections.

1)

VeM.va.:tion 06 :the Computa.:tion FoJUnu1.a.

Let us write our notations thus:

A=

i' = (1, .... 1)

(I - A)

= (I

= AB

- A

+ M)-l

= (0 10 2

= L o.

i=l

1-

....

o )'
n

56
C = the matrix of consumption

coefficients

_
n

= a~'a = ~ai=l
L na.
'l- 'l-

then we find
(3.16)
Since obviously 0

<

<

1, and we can prove 0

<

Hence, the inverse of (I - CH) surely exists.


(I - CH)D
CHD

n<

1, so that II - CHI = 1 - an

>

O.

Let us denote (I - CH)-l = D, then

D - I.

(3.17)

Multiply CH for each side of the above equation and we get

Because of (3.16), the left side of this equation is (CH)2D = anCHD, and hence
1

CHD = ~l
-- H.
- an

Substituting (3.17) for the left-hand side of the above expression, we obtain
1
-_H
- an

(3.18)

D=I+~l

Thus, taking into account (3.15) and (3.18), our final matrix multiplier is

[(r - c)(r - A)

M]-l = [I - A + M]-l[I - CH]-l

= B[I +

~l
-- H]

- an

(3.19)

For the sake of completeness, a similar formula for the case of a closed economy with
no foreign trade can be derived as:
[(I - C)(I - A)]-l

= [I - A]-l[I _ C]-l
=

1
B[I + .--'--=C
l - a-]

(3.20)

in exactly the same manner.

The above formula (3.19) also can be derived by tracing the propagation process
by the initial injections.

Denote by m the numerical order of the propagation and

57

let
X:
m

the increment of output of the mth order,

Y:

the increment of income (value added) of the mth order,

C :
m

the increment of consumption of the mth order,

then the income-generating process due to the initial injections f can be expressed
as
2'

1'

Y2 ---->

---->

3'
---->

Denoting by M the diagonal matrix of import coefficients, and by C the square matrix
of consumption coefficients, then
from 1
from l'

Cl = CY l

from 2

Y2 + MX 2

from 2'

C2 = CY 2

from 3

Y3 + MX 3

Cl

}
}

in general, the income increment of the mth stage is


m = 2, 3, ....

(3.2l )

On the other hand, the output increment of the mth stage is


X =AX +Y
m
m
m
Y

= {I - A)X =
m

Axm

(3.22)

Putting the above equation into (3.2l), we have


{,4 + M)X

ciix

x m = BCiixm- 1

m-

1
m = 2, 3, ....

(3.23)

We wri te the coefficient of the right-hand side of the above equation as BCA =
then
Xl = Bf
Xm

= QXm_l

m =

2, 3,

....

Q,

58

Hence

Xm

= ~-lXl = ~-lBf,

and since

= BCABB- l = B(CH)B- l ,

= B(CH)mB-l = B(an)m-l (CH)B- l

(an)m-1 Q

so that the output increment of the mth stage is


m

= 2,

3,

(3.24)

Thus

(3.25)
Since 0

<

an

<

1,
Bf + - 1
l_ii BeHf = B[I +.,.----'-=Cl
1__ HJf
(J
en

This result coincides perfectly with (3.19).

PART TWO
INTERNAL AND EXTERNAL MATRIX MULTIPLIERS
CHAPTER 4
INTERNAL AND EXTERNAL MATRIX MULTIPLIERS IN THE INPUT-OUTPUT MODEL *

I.

l~oduction

There are many useful applications in which the n industries of an input-output


table are divided into two or more strategic industry groups and the interaction
among these groups traced through.

To mention some examples, we have the interac-

tions between the goods-producing sectors and service sectors, between the primary
growth sectors and the supplementary or derived growth sectors, and between two
regions which have structurally different characters.

Another example is found in

the necessity of distinguishing industries subject to capacity limitations from those


which have ample capacity.

Examples of this sort might be given by the hundreds.

The method employed herein is to partition off the original Leontief inverse in
terms of the combined effects of "internal multipliers," "external multipliers," and
their "induced sub-matrix multipliers."

Such an "internal-and-external matrix

multiplier model" may well be applied to a wide variety of problems, as the usual
Leontief inverse conveys only the ultimate total effects of interindustry propagation
but not the disjoined effects separating into partial multipliers.
Part Two is an attempt to investigate problems of this type by means of the
formulation of partitioned matrix multipliers and their relationships, and to
empirically apply our formula in two cases:

1) the interdependent model of

goods-producing sectors and service sectors in the next chapter, and 2) interregional

* This and the next chapter are revised and integrated versions of articles which

were originally published, under the same titles, in Hito~uba6hi Jo~nal 06


Eeonomie Vol. 7 No.1, June 1966, and in the ame Jo~nal, Vol. 12 No.1, June
1971. The mathematical part of this chapter is a summarized version of Chap. 4
and Appendix C in the author's The Stnuet~ Int~dependenee Analyi 06 an
Eeonomy (in Japanese), 1963, and the empirical applications are from the author's
arti cl e (i n Japanese): "An Interregi ona 1 Input-Output Model and its Appl i cati on" ,
Monthty S~vey 06 Japan Indu.tJU.ai StJw.et~e Intitute. No. 52, Jan. 1965. I have
introduced some improvements in this rewriting.

60

repercussion model in this chapter.

II.

Pcur;tWoned Ma.tJU.x MuUipUeM

We divide the n industries of the usual input-ourput table into two subgroups,
designated P sector which consists of Z industries, and S sector which consists of m
Then, the n x n matrix of input coefficients is

industries.

Z + m =n

(4.1)

where P and Pl are submatrices of coefficients showing the input of P sector's


products in the P and S sectors respectively, and Sl and S are submatrices of coefficients showing the input of S sector's products in the P and S sectors respectively.
Among these submatrices,

and S are square having the orders Z x Z and m x m

respectively, Pl and Sl are retangular having the orders Z x m and m x Z respectively.


Since the n x n Leontief inverse
B*

= (I

- A) -1

(4.2)

tells us something about the total ultimate effects only but not about the disjoined
interdependence of the above two activities, we must introduce some device consisting
of partiti oned matri x multi pl i ers.

In order to' sol ve thi s problem. we decompose the

elements of the Leontief inverse into three aspects of propagation consisting of the
followi ng:
(i)

Internal propagation activities inside the P sector's industries. This

aspect will be shown as the "internal matrix multiplier" of the p sector (having
order Z x Z):
(4.3)

(ii)

Internal propagation activities inside the S sector's industries.

This

aspect will also be shown as the "internal matrix multiplier" of the S sector
(having order m x m):
T = (I - s)

-1

(4.4)

Of course, each internal matrix multiplier does not operate independently under its

61

own impetus, but is able to operate in conjunction with the other sector's industrial
activity.
(iii)

Intersectoral propagation activities between the P and S sectors'

industries. This aspect will be shown as four rectangular sub-matrix-multipliers


which naturally follow from the operation of internal multipliers Band T:
Bl = SlB ... S-goods input in P sector induced by internal propagation in

P sector's industries (m x t).


B2

= BP l

... internal propagation in P sector's industries induced by

P-goods input in S sector (t x m).


Tl

= P1T

... P-goods input in S sector induced by internal propagation in

S sector's industries (t x m).


T2

= TS l

... internal propagation in S sector's industries induced by

S-goods input in P sector (m x t).


These four sub-multipliers reveal the coefficients of induced effects on output
activities in the cases of B2 and T2 or input activities between two sectors in the
cases of Bl and Tl , and are ca 11 ed the "producti on-generati ng process in successi on."
(iv) The above three aspects of the interaction process naturally lead to
another intersectoral multiplier that we could call the "external matrix multipliers"
of the P and S sectors according to their economic meanings.

If we select the

coefficients of the induced effect on production (i.e., B2 and T2 ) as the base, then
they will take the form
(4.5)

or alternatively
(4.6)

Of course L, the external matrix multiplier of the P sector, has the order txt,
and K, the external matrix multiplier of the S sector, has the order m x m, because

62

the multiplications of rectangular matrices form new square matrices. l }


(v) Now then, we have arrived at the fact that the total of the propagation
effects in the P and S sectors' industries, each generated by its own sector's
activities, are expected to take the values

LB

and

respectively, i.e., "the

internal matrix multiplier" premultiplied by the "external matrix multiplier." So,


if we assume

LB

=N

then we can prove the following formula:


B*

or

= (I-A}-l

l
.[

B + Btml

B-!d

MEl

NTl

T2N

T + T~Tl

(4.7)

l.

In other words, we can break down the original Leontief inverse B*

(4.8)

= (I -

A}-l

in terms of the combined effects of internal and external matrix multipliers and
their induced sub-matrix-multipliers.

From which it is easily seen that the combined

effects in both P and S sectors, originated each in its own sector's activities, can
be written in the additive 6orom B + Btml or T + TzNT l as well as the muttpiied 6orom
1) Another formul ati on of the "external matri x multi pl i ers" based on the coeffi ci ents

of induced effect on intersectoral input activities (i.e., Tl and Bl ) could be

L = (I - T1Bl}-1

(Z x Z)

(4.5)'

(m x m)

(4.6) ,

and
where K has the order m x m, and L has the order Z x Z.
The existence of these inverses (external multipliers K, L, K and L) as well as
the existence of internal multipliers (B and T) is warranted by the existence of
the original Leontief inverse matrix.
For the existence of a meaningful solution for Leontief model and the properties
of Leontief-type matrices, see Chapter 1, Section IV, l}.

63

LB or KT. 2), 3)

In any case, such analysis serves to elucidate some inherent

properties of the interaction between the P and S sectors. 3)


The proof of the formula (4.7) is as follows:

I B,?!] [ I - P I -Pl ]
M L -Sl
I - S

[ B + Btml
MBl

= r~]
L~

B(I - p) + BfD31(I - p) - BzMSl

= I + B'?!SlB(I - p) - BzMSl

=I

= I + B'?!Sl - B'?!Sl

MBl (I - p) - MS l
= MS1B(I - p) - MS l
=

MS l - MS l =

2) Using the notation in note 1), we can prove the following identities:
KT=TX=M
LB=BL=N
that is, the internal multiplier potmutti~ed by the external multiplier is also
possible as well as the p~em~ptled expression.
The proof of the latter identity is that:
LB = (I - B2T2 )-lB = [I +

m=l

BL = B(I - T1Bl)-1 = B[I +

because

and

(B 2T2 )m]B = (I + B'?!Sl)B

m=l

(T1Bl)m] = B(I + P1MB1)

~ (B 2T2)m = ~ B2(T~2)m-1T2 = B2[ ~ (T~2)m]T2


.1.1
m~
= B2(I - T~2flT2 = B~T2 = B~Sl = B'?!Sl

Y(T1Bl)m = m=lY(P1TS1B)m

m=l

Pl(TS1BP1)m-1TS1B = Pl [ (T~2)m]TBl
m=l
m=O
= Pl(I - T2B2)-lTB l = P1KTB l = P1MB1
=

So, we obtain
LB = (I + BzMS1)B = B + BfD31
= B(I + P1MB1 ) = BL.
In exactly the same manner, we get
KT = XX = M.
3) For the equality between the multiplied and additive forms, see the equation (*)in
note 2) above.

64
- B2 - B2XT 2B2 + B2XT(I - S)
- B2 - B2XT 2B2 + B2X
- B2[I - X(I - T2B2 )]

=-

B2(I + XT 2B2 -

x)

=0

- MB1P l + M(I - S)
- XTB1P l + KT(I - S)

In exactly the same manner, we have

(vi) The divided matrix means that we use the following system:

x = PX
P

+ P1X + F
s
P

= SlXP

(4.9)

+ SX + F
s

where X is an output vector of P sector's industries, X is an output vector of S

sector's industries, and F ,F are the final demand vectors of the P and S sectors
p

respectively.

Thus the solution of this system is stated as

or

][~l.
The partitioned intersectoral activities may be viewed in two ways:

(4.10)
(a) The

first expression of the formula (4.10) shows it from the perspective of the P sector
and (b)

the second expression shows the same fact from the perspective of the S

sector.

These expressions go hand-in-hand to make the general formulation applicable

to a variety of problems.
One more alternative expression of the Leontief inverse in terms of partitioned
matrix multipliers is

65

B*

__ t_=LB_+-L__
BKTT-=-l:. . . ]
= (I - A)-l = --:::LB~+-L-::B:=2T_]
~KT_B

'"

KT

KTB1

4.11)

We can easily prove the identity between this expression and equation (4.10).4)
Mathematically, our formula also provides us with a method of working with lower
order matrix-calculations, which is especially useful when the inversion of matrices
of high order is not suitable for available computational equipment.

II 1.

1nteJc/teg..to n.ai. RepeJr.C1L6.6..to n. Model.

An empirical application of our model is made for the interregional-interindustry data of the Japanese economy. The main purpose of the interregional
input-output model developed by Isard, Leontief, Moses, Chenery and others,5) is to
analyze the interrelations among trade and production in two or more regions. Our
interna1-and-externa1 matrix multiplier model, in a somewhat extended form, may be
more useful for this purpose, because the inverse of the usual interregional
input-output model depict only the ultimate total effects but not the disjoined
effects separating into interdependence between each regions internal and external
multi p1 iers.
One example of an interregional input-output table is the data published by the
Hokkaido Development Bureau which divides the Japanese economy into two regions, each
having the same number of industries (30 industrial sectors).

In terms of our

formula (4.7), we let the P sector be Hokkaido and the S sector be the Rest of Japan.
For convenience we call the latter region Honshu, the main island of Japan. Of
course, Z = m = 30, and n = 60.
Table 1 is concerned with the internal and external matrix multipliers in each
4) For example, the identity KTt

is shown as follows:

= T2 (B + Btm1) = T8 1B + Tt~TB1 = (I + T2B~)TB1

T2N

in which I +

= T2N

Tt~

T2N

= K because

(I - Tt2)K

= I, so we obtain

= KTB1 = KT8 1B = KT2B.

5) See, W. Isard [22], W. W. Leontief [29], H. B. Chenery [5] [6], L. N. Moses [38].

66

region.

However only the column sum or row sum of the elements of the matrices are

cited because of limited space.

The column sum summarizes the pattern of "the power

of dispersion" of industries in each region, and the row sum measures "the sensitivity of dispersion" for industries in each region.
As shown by the figures in the Table, the internal propagation in Hokkaido (B)
has a multiplier effect of 1.77 on the average, and it evokes in turn a round-about
external repercussion through Honshu's industrial activity (L) of about 0.7%-up
effect on the average, so that the total effect is equal to 1.77 x 1.007
the average.

= 1.782

on

On the other hand, the internal multiplier effect in Honshu (T) is a

considerably higher 2.53, but the round-about external multiplier (x) shows only a
0.3%-up effect on the average.

This industrial differential reflected in the values

of the external multipliers L and X suggests the characteristics of each region's


industrial activity according to its role in the national economy.
To see this point more clearly, dividing the elements of internal multiplier of
Hokkaido

(30 x 30) by the elements of the appropriate (30 x 30) part in Leontief

inverse B* , we obtain the values which may called "inside propagation ratios" of
Hokkaido's industries. Although the table of the calculated figures is omitted, from
it we find that the most self-sufficing industries in Hokkaido are those in the light
industry group such as textiles, rubber products, leather and leather products,
printing and publishing, miscellaneous manufactures, and those in the non-manufacture
group such as services, trade, public utilities.
The industries in this category are relatively independent of Honshu's industrial
activity, and their "inside propagation ratios" all take values more than 0.9.

At

the other extreme, there is a group of industries highly dependent on Honshu's


industrial activity.

This group includes such heavy industries as iron and steel,

non-ferrous metal products, and the resourse industries such as metal mining,
non-metal mining, pulp, paper and paper products, fishing, etc.
Such internal propagation patterns, together with the external input patterns of
interindustry activity in each region, depict the characteristics of interregional
repercussions whose estimated results are summarized in Tables 2 and 3.

In these

tables the coefficients of interregional inducement relations are shown in a summary

67
TABLE 1
SUMMARY

TA~LE

OF INTERNAL AND EXTERNAL MULTIPLIERS OF

AN INTERREGIONAL MODEL OF THE JAPANESE ECONOMY


Internal
External
Internal
External
multiplier of multiplier of multiElier of multiElier of
Hokkaido, B Hokkaido, L Honshu T
Honshu K
row
Icolumn row
Icolumn row
Icolumn row
Icolumn
sum
Isum
sum
Isum
sum
Isum
sum
Isum
2.0757 1.5281 1.0053 1.00142.1371 2.0574 1.0016 1.0073
Public utilities
1.5973 1.5558 1.0103 1.0033 1.4232 1.8264 1.0041 1.0015
Metal mining
1.1008 1.5049 1.0010 1.0051 1.2693 1.4445 1.0005 1.0006
Non-metal mining
1.3339 1.9988 1.0013 1.0077 1.7521 2.2020 1.0017 1.0001
Petroleum and natural gas
2.5418 1.4247 1.0269 1.0040 1.7799 1.83841.0013 1.0007
Coal mining
1.7938 2.1285 1.0074 1.0017 2.5493 2.4767 1.0036 1.0025
Processed foods
Textiles
1.2630 1.5018 1.0012 1.00583.9697 2.9991 1.0069 1.0014
1.4638 2.0173 1.0021 1.0001 1.5687 2.7364 1.0006 1.0009
Saw-mill and plywood
Pulp, paper and products
2.1617 2.2074 1.0170 1.0037 3.1503 2.9812 1.0050 1.0056
Chemicals
1.84641.9450 1.0058 1.00484.2311 2.9287 1.0069 1.0026
Coal products
1.5500 2.2040 1.0073 1.0011 2.3138 2.4280 1.0052 1.0103
Rubber products
1.0272 1.5886 1.0000 1.0054 1.2967 2.0640 1.0014 1.0007
Leather and products
1.0242 1.2081 1.0000 1.0126 1.1281 2.9118 1.0001 1.0011
Nonmetallic mineral products 1.1340 1.8448 1.0007 1.0050 1.4833 2.4099 1.0015 1.0030
2.8235 2.7949 1.0626 1.0028 5.3194 3.8007 1.0163 1.0066
Iron and steel
1.0109 1.7924 1.0031 1.0093 2.0035 2.9756 1.0028 1.0108
Nonferrous metal products
Steel products
1.0922 1.3679 1.0005 1.0637 1.4305 3.0454 1.0027 1.0032
Machinery
1.0824 1.8586 1.0015 1.0267 1.8852 2.9986 1.0069 1.0010
Lumber and products
1.1370 1.9050 1.0008 1.0044 1.14082.5479 1.0002 1.0044
Printing and publishing
1.1219 2.0063 1.0003 1.0061 2.21982.7236 1.0016 1.0100
Miscellaneous manufactures 1.0409 1.7745 1.0000 1.0045 1.2726 2.9778 1.0015 1.0009
Forestry
2.5630 1.2480 1.0057 1.0008 3.4962 2.0360 1.0016 1.0001
Fishing
1.5319 1.9547 1.0026 1.0061 1.4503 2.2795 1.0002 1.0162
Agriculture
2.5916 1.4334 1.0097 1.0011 3.3701 1.7911 1.0047 1.0009
Dummy sector
2.1319 1.6158 1.0063 1.0113 4.8711 2.8317 1.0061 1.0007
Service
3.0585 1.4978 1.0046 1.00204.3531 1.8817 1.0030 1.0014
Business consumption
2.2541 2.4622 1.0047 1.0029 2.4325 3.1556 1.0016 1.0012
Trade
2.2341 1.4880 1.0069 1.00083.6191 1.9028 1.0031 1.0000
Transportation
2.6468 1.5602 1.0095 1.0046 2.1017 2.1303 1.0024 1.0017
Undistributed
2.8650 1.68181.0056 1.0018 4.7226 3.3584 1.0045 1.0018
(Average)

1.7699 1.7699 1.0070 1.0070 2.5247 2.5247 1.0033 1.0033

--------------------~------------------~--------------.----

68
form for both the column sums and the row sums of the four sub-multiplier matrices
B2 , T2 , Bl and Tl

Industries listed are only those having higher values than the

average.
The sub-multipliers B2

= BP l and

T2

= TS l are concerned with the propagation of

production activities in each region induced by the input activity in the other
region.

Reflecting the high dependence of Hokkaido's activities on Honshu's

industries, the elements of the multiplier T2 have higher values than those of the
multiplier B2 , with an average value of 0.3806 versus 0.0159. A similar situation is
found in the comparison between Bl

= SlB and

Tl

=P1T, which show the input induce-

ment effects of one region on the other, where the average value for Bl is 0.1847
versus 0.0207 for T l . These results suggest that a development program centered in
Hokkaido will give rise to many leakages in the interregional production process and
will generate much benefit to Honshu's industries.
We cite here one specific example of the nature of the regional industrial
interrelationship:

the production effect of Honshu's iron and steel industry induced

by Hokkaido's input is the extremely high row sum value of 2.303 as shown in Table
2-(a). This high value has its origin in Hokkaido's industries such as steel
products and machinery as shown the column sum figures in Table 2-(b). Of course,
to understand the cross-effects of this sort in detail, it is necessary to trace
back the source of the elements in the underlying matrices themselves instead of
simplying examining column sum or row sum values.
In any case, analyses of the above sort should playa role elucidating the
inherent properties of inter-and-intra regional industrial relationships, and we may
expect a fruitful application of this method (combining the extended model in the
next section) to the comprehensive data of the Japanese interregional input-output
table compiled by the Ministry of International Trade and Industry. This will enable
us to study the patterns of industrial interrelations among the nine regions of the
Japanese economy.6)
6) A report on this MITI 9 blocks-interregional input-output table [24] that took _
t~ree years of preparaEion_was published in 1967.
The nine regioDs are Hokkaido,
Tohoku, Hokuriku, Kanto, Tokai, Kinki, Chugoku, Shikoku and Kyushu.

69
TABLE 2
SOME COEFFICIENTS OF INDUCEMENT TO PRODUCTION
PER UNIT OF INPUT IN THE OTHER REGION *
(a) Row sum of elements of T2
Honshu's industry
Iron and steel
Textiles
Chemicals
Agriculture
Machinery
Metal mining
Leather and products
Coa1 products
Pulp, paper and products
(Average)

Hokkaido's industry
2.3033
.8631
.8006
.5751
.4907
.4317
.4097
.4016
.4016
.3806

(b) Column sum of elements of T2


Hokkaido's industry

Coal mining
Pulp, paper and products
Iron and steel
Forestry
Transportation
Fishings
Agri cu 1ture
Processed foods
Trade
(Average)

.0812
.0614
.0448
.0284
.0277
.0267
.0252
.0208
.0165
.0159

(d) Column sum of elements of B2


Honshu's industry

Steel products
2.2120
Leather and products
1.2278
Machinery
1.0868
Nonmetallic mineral products .8064
.6211
Textiles
Miscellaneous manufactures
.4637
Rubber products
.3936

(Average)

(c) Row sum of elements of B2

.3806

Printing and publishing


Coal products
Fishing
Iron and steel
Public utilities
Lumber products
Pulp, paper and products
Nonmetallic mineral products
Steel products
Processed foods

.0577
.0520
.0502
.0419
.0374
.0357
.0319
.0255
.0211
.0168

(Average)

.0159

* Sectors listed here are the industries having row sum or column sum values
higher than the average.
** Table (a) or (c) lists the names of the industry receiving the induced
effects, and (b) or (d) lists the names of the industry giving the induced effects.

70
TABLE 3
SOME COEFFICIENTS OF INDUCEMENT TO INPUT BY
INTERNAL PROPAGATION IN THE OTHER REGION*
(a) Row sum of elements of Bl

(Average)

Row sum of elements of Tl

Hokkaido's industry

Honshu's industry
Iron and steel
Machinery
Chemicals
Textiles
Leather and products
Processed foods
Metal mi ing
Agriculture
Coal products
Pulp. paper and products
Stee 1 products

(c)

1.0728
.5409
.4335
.4243
.3962
.3675
.3232
.2978
.2906
.2396
.2349

Coal mining
Pulp paper and products
Iron and steel
Processed foods
Transportation
Fishing
Agri cul ture
Nonmetallic mineral products

.1521
.1034
.0865
.0462
.0339
.0336
.0241
.0214

.1847

(Average)

.0207

(b) Column sum of elements of Bl


Hokkaido's industry

(d) Column sum of elements of Tl


Honshu's industry

Steel products
Nonmetallic mineral products
Leather and products
Machinery
Textiles
Miscellaneous manufactures
Fishing
Nonferrous metal products
Chemicals

.6243
.4534
.4517
.3721
.2629
.2387
.2354
.1960
.1951

Iron and steel


Coal products
Printing and publishing
Fishing
Pulp, paper and products
Public utilities
Nonmetallic mineral products
Steel products
Lumber products
Nonferrous metal products

.0474
.0451
.0427
.0410
.0368
.0353
.0320
.0303
.0275
.0212

(Average)

.1847

(Average)

.0207

* Sectors listed here are the industries having row sum or column sum values
higher than the average.
** Table (a) or (c) lists the names of the industry receiving the induced
effects, and (b) or (d) lists the names of the industry giving the induced effects.

71
IV.

Some Ex.:telt6ioY/-6

06

the Model

The model suggested in this chapter may be extended in two directions. The
first direction is an extension of the input-output model in an "intensive" manner.
The other direction is an extension of an "extensive" nature where the input-output
model is used in conjunction with various other models such as the macro-econometric
model, the linear programming model, and so forth.

In this section, we are again

concerned with intensive type extensions, and we limit our interest to the following
two cases.

1)

06

Ex.:teYl-6iort irt :the NumbeJt

PaJrA:.Ui..Orted GJLOUp.6

The line of reasoning which was employed in section II could be extended to a


larger number of partitioned industry groups (or regions), that is, the two-partition model can be expanded to three or more partitions.

However, the straightforward

procedure of dividing the model into a large number of industry-groups (or regions)
tends to degenerate into formalism, or is apt to be too complex.

We must not adhere

to too much form at the expense of the empirical spirit. 50, we prefer to employ
"a method of localizable partition" in which the redivision of groups is limited to
some particular strategic parts of industries (or regions).
The procedure of the extension employed here consists of two steps.
The first step is the re-partition of some "internal matrix multiplier",

step 1.

say the internal matrix multiplier of 5 sector, T = (I - 8)-1.

If we redivide the

5 sector into two sub-sections designated section 5-1 and section 5-2, it will lead
to the repartitioned matrix of input coefficients as follows:
p

sector
p

-~

A -

Uo
Vo

5 sector

f-ls::2
aO

vl

al
ul
v

)z
)j

)k

j +k = m

Z+m= n

(4.12)

'-..J' "----./ "----./

From the above coefficient matrix, we get the following relationships for the 5 sector's industrial activities:

72

i) Internal multipliers of section 5-1 and 5-2 in the 5 sector;

ii)

= (I

- u)

= (I

- v)

-1
-1

xk

(4.13)

(4.14 )

Intersectional (not intersectoral) sub-multipliers showing the inducement to


production between sections 5-1 and 5-2 inside the 5 sector;
Qu l

=a

j Xk

(4.15)

Rvl

= Il

(4.16)

iii) Localized external multipliers of section 5-1 and 5-2 inside the 5 sector (i.e.,
the meaning of "external" in this case is limited to within the 5 sector and
not beyond it);
U = (I - as)-l
V

= (I

- Sa)-l

(4.17)

(4.18)

k Xk

Then, according to the formula (4.11), we have the internal multiplier of the whole
5 sector in the redivided form, i.e.;
T

= (I

sf

-u l ]
[I = -Vl I I - v
U

-1

rLVSQ
UQ I UaR J
VR:J

(4.19)

step 2. The second procedure is the generalization of the above localized


multipliers into the relationships including the interaction between P and 5 sectors.
The pattern of this inter-and-intra sectoral relations is given by the following
Chart.
In this Chart, E, Q, R inside the three circles denote the internal multipliers
of the P sector, 5-1 section, 5-2 section respectively. L, U, V outside the circles
denote the external multipliers of P sector, 5-1 section, and 5-2 section respectively.
<P sector>

<5 sector>

section
5-1

section
5-2

73

With the aid of this Chart, we get six routes of inter-and-intra sectoral inducement relationships shown as follows:

(a)

(b)

(c)

<order of matrices>

<sub-multipliers>

<inducement routes>
5-2 -> 5-1

a.

= Qu l

j x k

5-1 -> 5-2

fl

= Rvl

P -> 5-1

a = Q(u O + U1Rv O)

xZ

P -> 5-2

, = R(V O + V1QuO)

k xZ

5-1

A = B(aOU + alVRv l )

Zx

->

5-2 -> P

Zxk

= B(alV + aOuQul)

The formation of these six sub-multipliers showing induced effects on production activity may be easily verified by tracing the repercussion routes between the above
sectors or sections in the Chart.
We have now arrived at a formula of the partitioned matrix multiplier, in this
case, which can be stated as follows:
[ LB

or
=

LAQ
LB
UaLB U(I + aLA)Q U(a. + aLIl)R
V,LB

L~R

V(fl + ,LA)Q

V(I +

,L~)R

(4.20)

The main course of the derivation of the formula (4.20) is that the system
X = PX + a_X + alX + F
P
P
(ru
v
p
{

Q)

+ FU

Xv = VrfP + vlxU + vXv + Fv

XU

= urfP + uKU + ulxv

(4.21 )

can be solved in a partiality form for the production level of the 5 sector (regarded as equations

and Q) ) by considering the economic causal succession on

routes of the induced effects.


The result is

74

[:: 1

Va

UQ

UaR

Vt

Vf3Q

VR

(4.22)

substituting this equation into (!) and collecting terms gives the formula (4.20).
By this formula, we see that the external (not localized) matrix multipliers of
sections S-l and S-2 are equal to UtI + aLA) and V(I + tL~) respectively.7)
There are two advantages in applying this formula to practical problems.

First,

the number of industries in each partitioned sector (or section) may not necessary be
the same (i.e., Z ~ j

k), and secondly the above treatment can be adapted for the

further subdivision of the particular part of strategic sectors or sections in succession, so we get a method of studying the various characters of industry groups
(or regions), say, according to their differing roles in the national economy.

2)

Inclu6ion

06 the. Inc.ome. FOJtmation

PJtoc.e,6~

The next extension of our model is the inclusion of the income generation process which is omitted in the usual input-output model. This omission is justified
only if the level of income and its use do not depend on the composition of production, because in this case a disaggregation of income generated by sector will add
nothing to an analysis of the aggregated Keynesian type.

But, under less rigid

assumptions this procedure is no longer valid, especially in the interregional model.


The location of production depends on the location of consumption, and the latter
cannot be determined separately from the calculation of the income generated in each
region.
In some studies, such as Chenery's Italian regional model ,B) the household sector
is transferred to the processing sectors from exogenous sectors and is regarded as an
industry whose output is labor and whose inputs are consumption goods.

But a more

7) An empirical application of our three-region model for Japanese economy is found


in Chap. 2 of Supplement Part of MITI [24].
8) See H. B. Chenery [5], or H. B. Chenery and P. G. Clark [7], Chap. B.

75

proper procedure in dealing with consumption is not to regard it as a fictitious


production activity, but to introduce the consumption function of a Keynesian type in
a disaggregated form. g ) As was shown in Chapters 1 and 3, this latter procedure
means, by implication, combining the Leontief propagation process and the KeynesianKalecki propagation process in a disaggregated form. 10 ) By such method the role of
income formation is taken out from the output determination in a separated and
exposed form which is indistinguishable in the usual standard models.

g) An empirical application of the three-region model that included the income


formation side is found in the above Report [24] noted in the footnote 7).
10) Some comments on Chap. 3 (K. Miyazawa [32]) for regional aspects are found in
R. Artle [1]. The discussion in Chap. 1 furnish a reply to the comments.

CHAPTER 5
AN ANALYSIS OF THE INTERDEPENDENCE
BETWEEN SERVICE AND GOODS-PRODUCING SECTORS*

I.

rntAoductio~

There are two opposing views regarding the relationship between the
goods-producing sector and the service sector, and this divergence of views has
some relevance in light of the recent growth of service sectors in some advanced
economies.

Under the orthodox view, service activities are secondary to

goods-producing activities; the income of the former is seen as derived from the
income of the latter, i.e., a "redistribution of income" originated with
goods-producing activities.
The distinction between "productive" and "unproductive" labors employed by Adam
Smith, and the concept of surplus value employed by Karl Marx, are some of the
earliest expressions of this orthodox view.

On the other hand, modern economics

treats service activities on an equal basis with goods-producing activities, since


many services yield utility as goods and thus have value in terms of the exchange
mechanism of the national economy.
Modern economics emphasize the demand aspects by its reliance on the concepts
of utility and exchange, whereas the orthodox view stresses the supply aspects by the
importance it attaches to the production-relationship. The contrast between the two
views is thus very clear:

modern economics taking a unified view of goods and

services, while the orthodox economics laying weight on the two-dimentional relation
in production and service activities.
These differences have been reflected in the oppos i ng "nati ona 1 economi c
accounting systems" of the present day.

The so-called SNA, i.e., the System of

* This paper appeared originally, under the same title, in

EQo~omiQ G~owth Centeh


ya.te UMVe1[,6Uy, No. 92, August 1970 and HUouubMYU JOM~a.t 06
EQOMmiM, Vol. 12 No.1, June 1971, and is reproduced here with some additions.
The author is indebted to Professor H. T. Patrick for his encouragement and advice
in preparing the original paper, and to A. S. Bha11a and Y. Shionoya for offering
valuable discussions.
VL6QUA-6-i.o~ Pa.p~,

77

National Accounts of the United Nations, employed for the economic accounts of many
capitalistic countries, includes both income from service activities and income from
goods-producing activities as part of the National Income.

On the other hand, the

so-called MPS, i.e., Material Product System, a system employed in socialistic


countries, does not include service income as part of the National Income or National
Product.
The one-dimensional approach to the income-concept adopted by SNA is superior in
the broad treatment of problems.

But if incomes from both goods-producing and

service activities have an internal relationship rather than a parallel one, we must
take this into consideration together with the modern economic view.
In this chapter we shall present some linkage for the gaps just described and
shall formulate the interdependent models of the goods-producing sector and the
service sector both by methods of income analysis and of input-output analysis with
some tentative empirical illustrations.

II.

lYlc.orne and Ernpioymertt AYlal!J,6-U, 06 lrtte.JtdepeYldeYlc.!J 06 Two SectaM

Denoting by Y the national income defined by the usual SNA concept, we divide Y
into two components:

Yp '

the income from the goods-producing activities, and Ys' the

income from the service activities, i.e.,


Y=Y

(5.1)

+Y

where Yp may be regarded as the national income defined by the MPS concept.

If we

define q as the propensity to consume service and ignore the intermediate services
for simplicity, then the demand for services, i.e., qY, must be equal to the supply
of services in equilibrium and we have
(5.2)
Substituting (5.1) in (5.2), we get
Y =~.y
s
1 - q P

(5.3)

This equation (5.3) demonstrates that the level of service-income Ys depends on two
factors:
services.

the activity level of goods-production and the propensity to consume

78
The first factor reflects the correlation of the level of service activities
with the level of goods-producing activities so that the latter determines the former as argued by the orthodox economists. As shown by the equation, service
activities will expand with a higher level of activities in goods production. The
second factor reflects the structure of demand as asserted by many modern economists.
Since the income-elasticity of demand for services is greater than that for goods,
the degree of growth in the service sector would depend on the order of increase in
the propensity to consume services. Higher propensity would generate a higher level
of service activities.
The form of expression (5.3) is very similar to the Keynesian multiplier
equation, and in effect we can interpret it to be the result of the propagation
process caused by the goods-producing activities through the expenditure of income.
Justification for this assertion is as follows.

First, we make three assumptions: (i)

the propensity to consume services q plus the propensity to consume goods


propensity to save
investment ex

po~t;

plus the

are equal to one (i.e., q + p + 8 = 1); (ii) saving equals

and (iii) the level of goods production equals Y ex

Given

po~t.

these assumptions, the following propagation processes can be worked out:


income from
goods-production

Y supply

of good~
II
demand .J
for goods

<total>

= _8_ y

savi ng

1 - q

consumption
of goods

:~/;~/~r:P
tit i t

consumption
of services
income from
service activities
national product

p]

=_p_y

Y=Y

+Y

=Y

+ P

--Y

l-qp

=Y

1 - q P

Y
demand for
1 - q pservices ~

= -fl-

2
3
-fl- Y
Y = qY + q Y + q Y +... = 1
8

-q P

supply of
services

II

+-fl-y =_l_y
1 - q p
1 - q P

Of course, the convergence conditions are generally satisfied by the assumption that
the value of q is less than one.
We could define the multiplier

as the "expansion multiplier of service

activities" accompanied by the multiplicand Yp' Thus formula (5.3) interpreted as a

79

multiplier equation may be one way of synthesizing the orthodox view of the relationship between goods and services with the modern Keynesian expenditure viewpoint.
In any event, the growth of the service sector must be explained in terms of both
demand and production.
The last line of the above table, i.e., the equation on national product,
y = ___1___ y

can also be obtained in another way, namely, by substituting (5.2) in

1 - q p'

(5.1).

This national product equation gives us a relationship between the national

income defined by SNA (i.e., Y) and the national income defined by MPS (i.e., Y ).
P

Some popular explanations for growth of the service sector in recent years are
often expressed in terms of employment rather than in terms of income.

Our formula

(5.3), expressed in terms of the income base, could be transformed into the employment base so that
L

--11- :.: L
1 - q YB P

(5.4 )

where LB ,Lp are levels of employment in the service sector and in the goods-producing
sector respectively, and
Y
YB-..lL
-L'
B

-...l2.

p-L

stand for the productivity of each sector.


According to A. S. Bhalla,l) attempts to explain the relatively faster growth of
employment in the service sector have followed three main approaches, namely:
income and expenditure approach, (ii)
approach.

productivity approach, and (iii)

(i)

employment

These three approaches are integrated in our formula (5.4).

The first element in (5.4), i.e. --11-1


,is a demand factor that reflects the
- q
value of propensity to consume services, and therefore the degree of income-elasticities in demand for services.

This aspect is emphasized by Colin Clark and others. 2)

The second element, yp /y B , could explain employment growth in services through a


1) A. S. Bhalla [3J.

2) C. Clark [lOJ.

80

relatively slower growth of productivity in the service sector, as pointed out mainly
by V. Fuchs. 3 ) And the third element,

L ,

shows that growth of service employment is

a function of the growth of manufacturing employment, as argued especially by W.


Galenson. 4 ) These conventional explanations of employment growth in the service
sector are often considered in isolation rather than in conjunction with each other,
as formula (5.4) now enables us to do.
Let us now return to formula (5.3), expressed in terms of incomes, in order to
examine its character and economic meanings.

At least four points should be noted

prior to any generalized treatment of the above formula.


(i)

First, in order to conceive of our formula (5.3) as one of the multiplier

equations, we have to assume that the component of the multiplier, i.e., q, is


independent of the multiplicand Y ; similarly, the multiplier formula, in general,
p

must adhere to this assumption.

However, the assumption does not hold for our case.

In modern capitalistic societies, the value of Y includes costs such as adverp

tisement and information, and the sum of these expenditures by firms influences the
value of q, or according to K. Galbraith, there are "dependence effects" or a shift
from "accepted sequence" to "revised sequence" in the relation between demand and
production in the market-place. 5) In a comment on Galbraith, R. Solow points out the
possibility that the effects of advertising by various firms may offset each other. 6 )
Still, we cannot overlook the existence of such effects.

At any rate, this problem

leads naturally to the next point.

3) V. Fuchs [13].
4) W. Galenson [15].
5) J. K. Ga 1bra ith [14], ch. [19].
"The unidirectional flow of instruction from consumer to market to producer may be
denoted the Accepted Sequence" (p. 221). " ... the accepted sequence is no longer
a description of the reality and is becoming ever less so. Instead the producing
firm reaches forward to control its markets and on beyond to manage the market
behavior and shape the social attitudes of those, ostensibly, that it serves. For
this we also need a name and it may appropriately be called The Revised Sequence"
(p. 212).
6) See R. M. Solow [48], p. 105.

81

(ii) The value of Yp in (5.3) reflects income

o~g~nating

in the goods-produc-

ing sector. To be sure, Yp is value-added by manufacturing industries, but not


neeeived in the same industries.

And Yp contains not only costs of advertisement and

information, but also interest costs, rent, transportation, insurance rates, etc.
These costs are transferred from the goods-producing sector to the "tertiary sector"
as income. The increasing trend of such income-transfer coincides with the fact that
growth in the goods-producing industries reflects the increased activities in the
service sector.
Then, denoting by

this transfer of income from the goods-producing sector to

the service sector, and by Y; income 4eeeived in the goods sector, we get
Y'

=Y

- T

(5.5)

Here Yp is income as defined by the MPS concept, whereas Y'p is the income of the
goods-producing sector in the usual sense employed by national accounts in capitalistic countries. Denoting by Y'8 the income of the service sector in the ordinary
sense, namely, on an income received basis, we have

= Y8

Y'8

+T

(5.6)

So, if we let
_ T
a - Y'

'

(5.?)

formula (5.3) would be rewritten in the following form:


Y'=~Y'
8

1 - q p

(5.8)

where a may be viewed as the ratio of transfer income.


This revised formula (5.8) is expressed in terms of an "income-received basis,"
and not an "income-originating basis" as was the case of the previous formula (5.3).
In this new expression, the fundamental assumption of independence between the components of the multiplier and the multiplicand may be supported.
parameter

An increase in

i.e., the ratio of transfer income, will induce an increase in Y',


namely,
8

82

the income of the service sector measured on the income-received basis. 7)


If we adhere to the income-received basis, the equation of the growth of service
employment must also be rewritten as follows:
L

=~
~
1 - q y'

(5.9)

where the definitions of productivity for each sector are revised as


y'
, _ s
Ys - L
s

'

y'
y'=-.R

L
P

respectively.
The tendency toward increase in the parameter a is illustrated by Table 1, which
is due to M. Shinohara. 8 ) Figures in Column (a) of Table 1 indicate the gross valueadded in manufactures according to Industry Census Statistics, where manufactures'
income is on the income-originating basis.

On the other hand, Column (b) reflects

the income of the manufacturing sector according to Income Statistics, where


manufactures' income is on the income-received basis. The trend of increasing
discrepancy in both sets of figures clearly discloses one reason for the growth of
the service sector in recent years. 9 )
It may be worth noting that the recent growth tendency of service sectors is
usually calculated on an income-received basis and not on an income-originating
basis.

Consequently, if we choose the latter rather than the former, namely (5.3)

instead of (5.8), or (5.4) instead of (5.9), the weight of goods-producing activities


would be larger than that which is usually estimated.
(iii) Third, the relationship between the growth of the service sector and the
activity level of the goods-producing sector (i.e., the formula (5.8)) suggests that
the latter determines the former, but this may represent only one side of the
problem.

7) Another expression of "the ratio of transfer income" may be written as S = T/Y ,


and we have a = S/(l - s).
P

8) See M. Shinohara [46J, ch. 3.


9) The value of a may be obtained by the expression a

fE+ - 1.

83
TABLE 1
DISCREPANCY IN MANUFACTURES' INCOME
BETWEEN INDUSTRY CENSUS AND INCOME STATISTICS

(a) Gross value-added


in manufactures
(by Industry Census)

(b) Income of
manufactures
(by Income Statistics)

-ffi-

hundred million yen

hundred million yen

1952

13,000

11 ,629

89.5

11.7

1953

16,864

13,092

77.6

28.8

1954

18,959

15,358

81.0

23.4

1955

20,986

15,348

73.1

36.7

1956

25,437

18,575

73.0

36.9

1957

29,522

24,082

81.6

22.5

1958

31,748

24,556

77 .3

29.3

1959

38,467

27,138

70.5

41.8

1960

50,348

36,390

72.3

38.3

1961

61,898

44,257

71.5

39.8

1962

71,505

49,815

69.1

44.7

1963

81,709

55,509

67.9

47.2

1964

94,619

63,935

67.6

47.9

Date

(l

* Source: M. Shinohara, Sa~gyokozo~o~ (1~dU6~ Stnu~e), Tokyo, 1966.


** Figures in column (a) show the manufactures' income in "income-originating

base," and those in column (b) show the manufactures' income in "income-received
base."
*** Column (l is an addition by the author.
(l = {(a)/(b)} - 1.

84
The other side, showing that service activities determine the level of
goods-producing activities, also exists, particularly if we take into account the
development of information-service industries in recent years.

The importance of

information-service industries is great enough that it is no longer possible to


regard them as merely an accessory of goods-producing activities.

Rather we must

treat information-service industries as a necessary and important part of the modern


economic activity.

Recognizing this characteristic, we leave the discussion on this

problem to future work lO ) and instead consider the twofold interaction between these
sectors from some other standpoint.
(iv)

Fourth, the approach to income analysis manifested in (5.3) and (5.8) has

one crucial limitation, namely, services are treated only as final products.

If

services did only constitute final products, then we could legitimately condust our
analysis in income terms.

It is true, however, that there has been a proliferation

of intermediate services as well as a growth in their magnitude in modern industrial


societies.

Thus the rising significance of intermediate services, which includes

the information activities, make it imperative that we examine the interactions


between final and intermediate services as well as between the goods and the
services sector. ll )
The next section presents a tentative approach to this problem based on an
input-output analysis.

Some of the points raised in this section will be more

adequately dealt with.


10) See the next footnote 11).
11) In the newest Japanese 1970 Input-Output Tabl~ published in 1974, an
enlargement of sector classification is introduced for the analytical purpose
on service activities. "The sector classification ..... could be used for the
input-output analyses on the recent radical change of the industrial structure,
especially on the expansion of service industries, on so-called system
industries such as information industries, housing industries. The increased
numbers of row sectors and column sectors were about 70 and about 60
respectively comparing with the [465 x 341] sector classification for 1965."
(the Repo~, English edition, p. 62).
We may be expected a fruitful application of the method developed in the next
section to this comprehensive data.

85

Inpu,t-OtLtpu,t Anallj.6-w 06 the Intvr..dependenclj 06 Two SedaM

III.

As an extension of the conventional input-output analysis, we have introduced in


Chap. 4 a formula of partitioned matrix multipliers depicting the interaction among
two or more strategic industry groups.

The method employed therein is to partition

off the original Leontief inverse in terms of the combined effects of ninternal
multipliers," "external multipliers," and their "induced sub-matrix multipliers."
Such an "internal-and-external matrix multiplier model" is appropriate for our present problems, as the usual Leontief inverse conveys only the ultimate total effects
but not the disjoined effects separating into partial multipliers.

At the risk of

being too repetitive, we summarize the results of our theoretical ideas and then
discuss the results of two empirical studies dealing with the interaction between the
goods-producing and the service sectors.

1)

I nteM edoJta.t PM pag alia n Pa:ttVtYl

We divide the n industries of the usual input-output table into two subgroups
designated P sector, i.e. goods-producing sector which consists of Z industries, and
S sector, i.e. service sector which consists of m industries.

The segregated input-

output system is expressed as


=

PX + P1X + F
P

= SlX

SX

(5.10)

8,

where X ,X are output vectors and F ,F are the final demand vectors for the P and
p

S sectors respectively.
ways:

The separate intersectoral activities may be viewed in two

(a) the first expression of the formula is from the P sector's viewpoint and

(b) the second expression constitutes the S sectors's viewpoint.

As shown in Chap. 4,

the solution of the system is stated as

where

= LB,

t;: 1

[T;N T:T~2NT j

= KT, in which

= (I

prl and

(5.11)

= (I

sr l are "internal multi-

86

and S sectors respectively.

Four rectanglar sub-matrix multipliers showing the

intersectoral propagation activities between the P and the S sectors' industries, are

The first empirical application of our model utilizes the interindustry data,
consisting of 50 goods-producing industries and 4 service industries, of the Japanese
economy published by the Japanese Government under the cooperation of the Economic
Planning Agency and five other Ministries (the t1inistry of Agriculture and Forestry,
the Ministry of International Trade and Industry, the t1inistry of Construction, the
Statistics Bureau of the Prime
Agency).

t~inister's

Office, and the Administrative Management

Utilizing formula (5.11), we let the P sector represent the 50 goods-pro-

ducing industries and S sector the 4 service industries, i.e., l = 50 and m = 4.


Oividing the elements of the internal matrix multiplier of the goods-producing
sectors

(50 x 50), calculated from the above equation, by the elements of the

appropriate (50 x 50) part in the published Leontief inverse B* = (I - A)-l, we


obtain the values which may be called "the inside propagation ratio of goods-producing sectors".

The extent of an industry's dependence on service activity can be

ascertained by examining the row elements of the 50 x 50 array of computed ratios.


Those industries having many higher value ratios are the less service-dependent
industries, and vic.e. ve.Ma.
Table 2 is a summary version of this empirical test, and shows industry-categories of goods-producing industries by degree of dependence on service activity.
Those in category A have characteristics which make them relatively independent of
service activity, while those in category D are at the other extreme.

Roughly

speaking, catogories from A-l to D-2 may be thought of as successive gradients of


dependency on the service sector.
In group A, the "inside propagation ratios" of each industry take values predominantly more than 0.9 (in A-l group), or more than 0.8
Those in Group B have ratios in the range of 0.7
spread far and wide in the range of about 0.5

0.9.

0.9 (in A-2, A-3 groups).

In Group C, the ratios are

0.9, and among this category the

87
TABLE 2
INDUSTRY GROUPS BY DEGREE OF DEFENDENCE ON SERVICE ACTIVITY
Groups

Names of Goods-producing Industries


Basic chemicals, Non-metallic minerals.

Group A

2 Electricity, Intermediate chemicals, Pig iron, ferro-alloys and


crude steel, Metallic ores, Non-metallic mineral products.
3 Rolled steel, Natural fibre yarns, Coal and lignite, Non-ferrous
metal ingots, Chemical fibre yarns, Metal products, Forestry,
Coal products.

Group B

Group C

Group D

Machinery and instruments (except electric), Steel casting and


forging, Miscellaneous crops.
2 Primary non-ferrous metal products, Saw-mills and plywood,
Chemical fertilizers, Fabrics, Rubber products, Pulp.
Leather and leather products, Livestock, Furniture and wood
products, Rice, wheat and barley, Electric machinery and
equipment.
2 Starch, sugar, seasonings, etc., Miscellaneous textile products,
Crude petroleum and natural gas, Paper and paper products,
Miscellaneous processed foods, City gas and water services,
Repair and maintenance of machines, buildings and structures,
Petroleum products.
Rice and barley polishing and grain-flour mills, Miscellaneous
manufactures, Fisheries, Printing and publishing.
2 Drugs, soap and cosmetics, Transport equipment, Manufactured
tobacco and beverages.

* Based on 1955 input-output tables. Goods-producing sectors listed here


excludes the dummy industries such as Business consumption, Office supplies,
Scraps, and Undistributed.
** Service sectors other than the above goods-producing sectors are Wholesale
and retail trade, Transportation and communication, Real estate and ownership of
dwellings, and Banking, insurance and services.
*** The order of listing is that the industries in Group A are the most
service-independent sectors, and those in Group D are the most service-dependent
sectors.
ratios in the C-2 group concentrate in the 0.7

0.8 range.

propagation ratios take lower values ranging from about 0.4

In Group D, the inside


~

0.5 to 0.7, and the

industries in this group are the most service-dependent ones.


The rationale for the above industrial differential-pattern may be traced to
the difference between the values of the elements in the 50 x 50 part of the

88

Leontief inverse B* and the values of the elements in the internal matrix multiplier
B, which is equal to BzMB 1 as is shown in the formula.

So, we must determine the

relative weight of B1 , B2 and M in the propagation process.


By turning our attention to Tables 3 and 4, we can ascertain what goods-producing industries have more inducible power for service activity (see Table of values
for B1 ) and what service industries have more inducible power for goods-producing
activity (see Table of values for B2).
particular interest:

A general feature of the Tables is of

an overall comparison of the estimated values of these two

intersectora1 sub-multipliers suggests that the weight of B2 is smaller than that of


B1 in propagation activity.

In Table 4 (B 2 ), values of more than 3% number less than

could be counted on the fingers of both hands (excluding the Undistributed sector),
while in Table 3 (B 1 ), values of more than 3% are numerous.

In other words, the

inducible power of one sector to another is more powerful in the case of the goodsproducing sector than in the case of the service sector.

Needless to say, there are

differential effects from one industry to another industry as is evident in the


Tab1es. 12 )
Of course, from the viewpoint of the goods-producing sector, the sub-multiplier
B1 operates on that sector only in an indirect manner in the sense that it needs a

medium operator expressed by M = KT as shown by the equation (5.11).


elements of

K,~nd T

The values of

are summarized in the Table 5 which shows the powers of disper-

sion of service sectors internally and externally.

On the whole, many values of the

elements in the internal multiplier T are somewhat higher than those in the external
multiplier K (except Real estate's column), but the difference between the values of
these two multipliers is not so large.

This fact means, again, that the weight of

12) One comment is needed because of the weakness in the data of the service sector
which leads to the estimation errors in the original Leontief inverse matrix. If
this data weakness is not negligible, our method must be reread in such a way
that the proportion of errors in the elements of the Leontief inverse is actually
due to a shortcoming of the service sector's data. For example, the reliability
of the inverse-elements may be judged by means of Table 1 such that those in
Group A-l are the most reliable and those in Group D-2 are the most unreliable.

89
TABLE 3
COEFFICIENTS OF SERVICE-INPUT INDUCED BY INTERNAL PROPAGATION IN GOODS-PRODUCING SECTOR*
(unit: 10 -6 )
B'l = (81B)'
TranSJX)rtaBanking,
tion and Real
Insurance
Sector
Trade Communi
ca- estate and
tion
Servi ces
28382
10876
12243
Rice, wheat and barley
347
18595
15910
50790
614
Miscellaneous crops
61239
27121
24374
897
Livestock
13509
10888
1830
23862
Forestry
54357
21559
6341
40338
Fi sheri es
22829
31772
Coal and lignite
32782
4228
25045
39486
Crude petroleum and natural gas
30837
27628
23816
20216
26165
8939
Metallic ores
21651
15567
20540
4514
Non-metallic minerals
14102
Rice and barley polishing and grain-flour mills 55809
16636
755
Starch, sugar, seasonings, etc.
116074
26157
2085
32663
Manufactured tobacco and beverages
33969
15115
767
19457
Miscellaneous processed foods
38199
40303
95402
2984
Natural fibre yarns
26100
28528
16786
1944
Chemical fibre yarns
52525
51397
5743
50083
Fabrics
34169
55940
5023
40814
Miscellaneous textile products
71816
32592
4105
46784
Saw-mills and plywood
18568
15909
2841
34709
Furniture and wood products
37096
39130
3036
49755
Pulp
32338
27412
35650
3631
Paper and paper products
46449
45912
2881
36683
Printing and publishing
52511
63314
87273
2326
Coal products
27008 147083
27438
3056
Petroleum products
11974
17553
10508
7081
Basic chemicals
40572
96450
4175
35066
Chemical fertilizers
39844
45497
67430
2948
Intermediate chemicals
83199
48159
56255
5579
Drugs, soap and cosmetics
82932
44149
5963
128804
Rubber products
39029
27059
3449
37856
Leather and leather products
39011
17434
26308
2685
Non-metallic mineral products
66153
66697
4087
36925
Pig iron, ferro-alloys, crude steel
24252
57757
16368
2506
Steel casting and forging
41869
46669
3209
26740
Rolled steel
29437
3383
23905
56458
Non-ferrous metal ingots
26865
38975
15111
5811
Primary non-ferrous metal products
30039
37476
4037
21291
Metal products
44686
42068
2468
26247
Machinery and instruments
56789
37272
6127
34695
Electric machinery and equipment
74690
38973
34070
2560
Transport equipment
76278
41917
2686
29493
Repair and maintenance of machinery, etc.
70881
46947
3204
25575
Miscellaneous manufactures
73958
37369
3834
47743
Electricity
23340
51460
6996
29435
City gas and water services
38028
40696
54657
2783
Business consumption expenditure
65213 236774
453851
816
Building construction
69495
59666
3624
33835
Miscellaneous construction
53675
68526
3734
37094
Office supplies
145474
56522
3146
57429
Scraps
8902
12466
612
4975
Undistributed
60336
36822
15166
68993

* This Table is shown in transposed form, interchanging rows and columns of


the matrix B1 for convenience.

90

TABLE 4
COEFFICIENTS OF INTERNAL PROPAGATION IN GOODS-PRODUCING
SECTOR INDUCED BY INPUT IN SERVICE SECTOR

Sector
Rice, wheat and barley
Miscellaneous crops
Livestock
Forestry
Fisheries
Coal and lignite
Crude petroleum and natural gas
Metallic ores
Non-metallic minerals
Rice and barley poiishing and grain-flour mills
Starch, sugar, seasonings, etc.
Manufactured tobacco and beverages
Miscellaneous processed foods
Natural fibre yarns
Chemical fibre yarns
Fabrics
Miscellaneous textile products
Saw-mills and plywood
Furniture and wood products
Pulp
Paper and paper products
Printing and publishing
Coal products
Petroleum products
Basic chemicals
Chemical fertilizers
Intermediate chemicals
Drugs, soap and cosmetics
Rubber products
Leather and leather products
Non-metallic mineral products
Pig iron, ferro-alloys, crude steel
Steel casting and forging
Roll ed steel
Non-ferrous metal ingots
Primary non-ferrous metal products
Metal products
Machinery and instruments
Electric machinery and equipment
Transport equipment
Repair and maintenance of machinery, etc.
Miscellaneous manufactures
El ectri ci ty
City gas and water services
Business consumption expenditure
Building construction
Miscellaneous construction
Office supplies
Scraps
Undistributed

TrCll5portation and
Trade Communication
3089
2242
3990
4058
1169
1079
11461
15325
2103
1156
3005
29780
73
1616
430
1642
692
2136
2725
1959
3396
3126
7770
1782
7226
8824
3240
4244
1420
1770
4038
4231
3241
4515
11377
17033
7463
3619
7519
3273
27389
11564
29721
10536
1142
13318
1739
46832
1608
2455
579
578
5821
7316
4200
1521
1497
8185
492
445
2238
7239
3897
15750
1117
4956
4982
20187
1038
3919
1621
7347
3532
7256
1924
7657
1800
4128
539
15919
12526
51220
5627
1773
6751
15826
1427
2142
81569
18212
0
0
0
0
37665
4861
3396
7789
45826
82680

Banki ng,
Real Insurance
estate and
Services
29239
347
19022
4637
14622
571
10104
12491
417
18837
8463
3805
152
60
2057
938
1191
2169
28287
582
13335
754
39452
708
51582
2185
1263
4274
1610
374
4434
912
4332
465
15608
5156
1387
2959
4573
749
16491
2640
28052
2566
2857
4028
1743
3608
1079
4008
3375
465
4482
18662
462
21077
2142
4592
215
1204
16036
5135
19479
7558
2441
12340
22222
9382
4924
2325
2412
8507
17153
6619
15969
5234
11355
1660
4408
9116
168849
13629
1358
5097
3510
14921
544
6594
41333
7124
0
0
0
0
1496
8926
9057
6812
16110
42035

91
TABLE 5
INTERNAL AND EXTERNAL MULTIPLIERS IN SERVICE SECTOR
(1)

Internal Multiplier of Service Sector:


Trade

= {I

- 8)-1

Transportation
and
Communication

Real
estate

Banking,
Insurance
and Services

1.006382

14208

1618

40780

Transportation and communication

49969

1.020766

1205

30004

Real estate

20839

4942

1 .000402

12839

Banking, insurance and services

61886

54474

31979

1.042842

Trade (wholesale and retail)

(2)

External Multiplier of Service Sector: K = {I - TzB2)-1

Trade
Trade (wholesale and retail)

Transportation
and
Communication

Real
estate

Banking,
Insurance
and Services

1.018500

14691

13090

21801

28305

1.015694

10241

21579

2004

2449

1 .000991

2099

49081

20060

8373

1.034038

Transportation and communication


Real estate
Banking, insurance and services

dependence of the service sector on the goods-producing sector is considerably large


in nature.

2)

CO.6;t- Pw.,h E66ec..t6 06 Sef1.vic.e- PJUc.e.6

We now switch our orientation from quantity-determination to price-determination


and use our input-output system to study the cost-push effects of service-prices on
the prices of the P sector's products.
Obviously, the prices of the P sector's products are given by the following
equations:
P

= P 'p

+ 8 'p

1 s

(5.12)

where pp ,p8 are vectors of prices of the P sector's products and the S sector's

92

service-outputs respectively, v is the P sector's vector of values-added per unit of


p

output, and tne coefficient matrices p' and Sl' are transposes of the matrices p and
Sl in the quantity model.
This price formation equation system (5.12) is a part of the following larger
model:

1
p

= p' pP

+ Sl 'p + v
8
P

= P l 'pP + s' p8

(5.13)

+v .
S

In this system, we take p and v as exogenously determined,and p and v as dependent


8
p
P
S
variables. The variations of p are due to cost-push effects, and, if we wish, the
p

variations of v could be viewed as the resultant change in wages or profits in the S


8

sector due to the variation in prices of the P sector's products.

But here we omit

this latter relation. Of course, the selection of endogenous and exogenous variables
is dependant upon the nature of the problem which we set forth.
The vector of determined prices in the goods-producing sector is represented by
the equation:
p
p

(I - p'rl{Sl'p

=B'{S'p

+ v }

+v}

p'

(5.14)

where B' is the transpose of the internal matrix multiplier of the P sector in the
quantity model.
If service-prices rise from p 8 to p8 + dp S , the resultant price-increases in the
P sector wi 11 be
dp
P

= B' S1
'dp = (s B)' dp = B 'dP .
81818

(5.15 )

Thus, in order to determine the cost-push effects of a rise in the levels of service
prices we need to transpose the sub-matrix multiplier Bl of the quantity system.
Returning to Table 3 and examining it from the perspective of cost-push effects,
we discover that:

(a) relatively more stimulated effects are brought on by a rise

in the prices of the Trade industry and of the Transportation and communication
industry than by a rise in the prices of the other service sectors, and (b)

the

93

resultant cost-push effects tend to be concentrated in some particular commodities


such as Starch, sugar, seasonings and

~1iscellaneous

processed foods in the case of

increases in Trade service-cost, and in Coal products and Basic chemicals in the case
of a rise in Transportation and communication cost.
Some caution is required in the evaluation of Table 3 from the perspective of
cost-push effects.

The estimated coefficient values may tend to be higher than the

realistic ones, because the constant input coefficients are predicated on the absence
of variation in relative prices.

Any changes in relative prices evoke input

substitution effects which tend to set limits on price rises of the cost-push type.
On the other hand, the recent rising trend in the service-input coefficients of
Japanese industries leads to an underestimate of the actual coefficient values of
Bl"

because the Table is based on somewhat old data.

IV.

Int~~ona Comp~on

The second empirical application of our model pertains to input-output tables


of several countries.

The data arrangements and calculations were done by staff

members of the Economic Planning Agency of Japanese Government.


include:

(1)

Input-output data

Japanes 1960 table by the Japan Statistics Bureau of the Prime

Minister's Office (jointly with the other government agencies), (2)


by the U.S. Department of Commerce, and (3)

U.S. 1958 table

West German 1960, French 1959, Italian

1959, Dutch 1959 and Belgian 1959 tables by the Statistical Office of European

Economic Community.

The number of sectors are uniformly aggregated to 34 from the

original classifications of each country's table.


Table 6 summarizes the nature of the interaction between the P and the S sectors
of each country in terms of the relation between the cross-input-coefficients of the
two sectors (i.e. Pl and 81 ) and the internal propagation of the goods-producing
sector (i.e. B).

By examining Table 6-(1), we can determine the extent to which a

country's goods-producing sector generates service activity, as the magnitudes in


this table are the row-sums of the coefficients of service-input generated by the
internal propagation of goods-producing sectors, i.e. the sum of the values of
Bl

= 8 1B for each service sector.

94
TABLE 6
INTERNAL MULTIPLIERS IN THE GOODS-PRODUCING SECTOR AND
LEVEL OF SERVICE ACTIVITIES
(1)

Coefficients of Service-Input Induced by Internal Propagation


in the Goods-Producing Sector
Country

Category

United Japan West


Germany France
States

Trade
Banking and Insurance
Real Estate
Transportation
Communication
Public Services
Other Services

D.0778
0.0130
0.0242
0.0490
0.0053
0.0135
0.0549

All Servi ce Sectors

0.2377 0.1675 0.1611

Italy Holl and

0.0687 0.0816 0.0300 0.0231


0.0235 0.0159 0.0197 0.0333
0.0005 0.0000 0.0000 0.0000
0.0480} 0.0474 0.0297 0.0340
0.0100
0.0054 0.0068
0.0008 0.0078 0.0011 0.0000
0.0159 0.0084 0.0356 0.0125

0.0496
0.0176
0.0000
0.0116
0.0072
0.0029
0.0324

Belgium
0.0456
0.0158
0.0000
0.0354
0.0081
0.0000
0.0185

0.1216 0.1097 0.1216 0.1234

* Figures in this table are obtained by summing up the elements of B = S B


for each service sector (i.e., figures are the ~ow ~um values of the eleme~ts of
Bl ), and indicate the coefficients of service-input generated by the internal
ptopagation of goods-producing sector.
** The names of the sector listed here are industries ~ec~v~ng the induce
effect.
(2)

Coefficients of Internal Propagation in Goods-Producing Sector


Induced by Goods-Input in Service Sector

Category

Country

United Japan West


Germany
States

France

Italy Holl and

Belgium

Trade
Banking and Insurance
Real Estate
Transportation
Communication
Public Services
Other Services

0.1526
0.0721
0.2719
0.2979
0.1242
0.1470
0.5461

0.2346} 0.2507
0.4507
0.1944 0.2320 0.3330 0.1501

All Service Sectors

0.2211

0.3162 0.2167 0.1338 0.1710 0.3057 0.1908

0.2403 0.2365 0.1425 0.1699


0.1937 0.1165 0.0846 0.0737
0.3269 0.2874 0.0000 0.2125
0.5873} 0.3375 0.4114 0.3979
0.3034
0.1206 0.1490

0.2841
0.1588
0.4244
0.6082
0.1868

0.2886
0.1150
0.1480
0.5108
0.0931

* Figures in this table are obtained by summing up the elements of B = BP


for each service sector (i.e., figures are the column ~um values of the el~mentslof
B?), and indicate the coefficients of internal propagation in the goods-producing
sector induced by goods-input in the service sector.
** The names of the sector listed here are industries giving the induced
effects.

95

One feature of the computed results is of particular interest.

Of the over-all

values in the last line of Table 6-(1), that of the United States is distinctly high.
It totals 0.24, while values for the other countries are in the range of about 0.11
0.17.

Among these other countries, Japan and West Germany have relatively higher
Thus, the capacity of the industrial sector to induce service activity is

values.

greatest in the United States, followed by Japan and West Germany.

Now, looking at

the table by service category, we find that the Trade category has a relatively high
value in almost all of the countries cited above, but, at the same time, it may be
worth noting that the Other Services category of the United States also has a significantly high value.

This shows that goods-producing activities, especially in the

United States, have a significant effect upon the Other Services category as well as
on Trade.
Table 6-(2) tells us what sort of service activity has more influence on the
internal propagation of goods-producing activities, because the numerical values are
the sums of the coefficients of internal propagation in the goods-producing sector
induced by goods-input in the service sector, i.e. the column-sum of the coefficients
of B2 = BP l for each service sector.
Looking at the table by category, the values for Transportation are highest in
every country except for the United States. Thus transportation appears to have the
greatest capacity to induce goods-producing activities.

In the United States alone,

the highest value is found in the Other Service category, and this fact, together
with the above conclusions, suggests that the United States is a country which has
the most advanced "Service Economy".

Considering the over-all values in the last

line of Table 6-(2), Japan has the highest value of .31.

This situation is due to

the relatively high value of the internal multiplier of the goods-producing sector
in Japan (on the average, the value for Japan amounts to 2.298, while that of the
United States is 1.843 and that of West Germany is 1.732).

On the other hand,

Japan's over-all value in the previous Table 6-(1) was not the highest because of the
relatively low service-input coefficients in the goods-producing sector (81) in
Japan.

On the average, the value of service-input in the P sector of Japan amounts

to 7.3%, while that of the United States is 12.9% and that of West Germany is 9.3%.

96
TABLE 7
INTERNAL AND EXTERNAL MULTIPLIERS IN SERVICE SECTOR
(1 ) Internal Multiplier of Service Sector
a)

Column sum values


Country

United Japan West


States
Germany France

Italy Holl and Bel gium

Trade

1 .2362 1.1459 1 .1071

1.1801

1.0907 1.2479 1.0433

Banking and Insurance

1.5427 1.2196 1.1001

1.0640 1.0485 1 .1358 1 .0599

Real Estate

1.1578 1.0160 1.2146 1 .0393 1.0520 1. 1019 1.0177

Transportation

1.2423 1.0192 1. 0979 1.2079 1.1589 1 .2175 1.0684

Communication

1.0873 1.0521

Public Services

1.0987 1.0479 1.0000 1.0000 1.0000 1.0000 1.0000

Other Services

1.3199 1.1788 1.1050 1.0851

Category

b)

1.0000 1. 1744 1. 1656 1 .0701

1.0655

1.0426 1.0929 1 .0514

Row sum values

Trade

1.1374 1. 1081

Banking and Insurance

1. 3981

Real Estate

1.3192 1.0129 1.0092 1.0000 1.0000 1.0000 1.0000

Transportation

1. 1816

Communication

1.0799 1.0823 1.0000 1.0588 1.0490 1.1052 1 .0665

Public Services

1. 1315

Other services

1.4363 1.1559 1.1170 1. 1132 1.1306 1. 1735 1.0587

1.0539 1.0535 1 .0501

1 .0276 1.0175

1 .2427 1.2067 1.1788 1.1469 1.0952 1. 0548


1.1408 1.1503 1.3309 1. 1817 1.3327 1. 1047
1.0088 1.0876 1.0156 1.0000 1.0419 1.0000

* Figures in a) are calculated as the column sum values of the elements of the
matrix T, and figures in b) as row sum values of the elements of the same matrix.
** Table a) lists the names of industry g~v~ng the induced effects, and b)
lists the names of industry ~eceiv~ng the induced effects.

97

(2)

External Multiplier of Service Sector


a)

Column sum values


Country

United
West
States Japan Germany France

Italy Holland Belgium

Trade

1.0268

1.0202

1.0250

1.0121

1.0138

1.0253

1 .0159

Banking and Insurance

1.0134

1.0248

1.0124

1 .0092

1. 0062

1.0157

1. 0066

Real Estate

1.0519

1.0339

1. 0271

1.0000

1 .0189

1.0355

1 .0117

Transportation

1.0501

1.0568

1 .0350

1.0339

1.0309

1.0416

1. 0289

Communication

1.0234

1.0379

1.0000

1 .0103

1 .0125

1.0150

1.0059

Public Services

1.0275

1.0245

1.0000

1.0000

1.0000

1.0000

1.0000

Other Services

1.0943

1.0467

1.0270

1 .0181

1 .0195

1.0270

1.0092

Category

b)

Row sum values

Trade

1.0718

1 .0705

1.0535

1 .0130

1 .0161

1.0507

1.0222

Banking and Insurance

1 .0211

1.0557

1 .0140

1.0128

1.0343

1.0198

1 .0125

Real Estate

1.0379

1 .0011

1.0000

1.0000

1.0000

1.0000

1.0000

Transportation

1.0510

1.0710

1.0433

1.0313

1. 0315

1.0246

1 .0266

Communication

1.0090

1.0237

1 .0000

1.0050

1.0074

1.0130

1.0069

Public Services

1. 0226

1 .0014

1.0079

1.0009

1.0000

1.0056

1.0000

Other Servi ces

1.0740

1 .0215

1.0078

1.0206

1.0125

1.0464

1.0100

* Figures in a) are calculated as the column sum values of the elements of the
matrix K, and figures in b) as row sum values of the elements of the same matrix.
** Table a) lists the names of industry giving the induced effects, and b)
lists the names of industry ~ecelving the induced effects.

98

Focusing on the service sector of each country, we find that the sUb-matrixmultipliers B1 and B2 operate on that sector in an indirect manner. Of course, the
final propagation effect on the service sector itself may be shown by the value of
M or by the separate values of T and K, namely, the internal and external multipliers

of the service sector (M

= KT).

The summary values of the elements of T and K are

given in Table 7. These values reflect the power (or the sensitivity) of dispersion
of the service sector both internally and externally.
Table 7-(1) conveys the effects of internal propagation on the service sector
starting from service-inputs in the service sector itself.

By examing the values

for each category in that Table 7-(1)-a), it is apparent that the service sector of
the United States has the greatest internal propagation effects. The values for
Japan are of about the same order as those for the European countries.

Such internal

propagation in the service sector leads, in turn, to circular repercussions on the


service sector itself through the goods-producing activities that start with the
consumption of goods in the service sector.
For example, the United States' Trade category has an internal multiplier effect
of 1.2362 on the average, and it leads to external repercussions through goodsproducing activities to the extent of an approximate 7.18%-plus average. We see then
that the total effect on the Trade category is equal to approximately 1.2362 x 1.0718

= 1.3250 on the average. Such is the manner in which the internal propagation
patterns together with the external repercussion patterns determine the characteristics of intersectora1 propagation in the service sector.
Among the round-about external effects in Table 7-(2)-a), the Transportation
category has the highest values in all countries except the United States.

Again,

the United States alone has its highest value in the Other Service category. This
fact reinforces our earlier conclusion that the United States of America has the
most advanced Service Economy from the point-of-view of the interaction of
goods-producing activities and service activities.
A comment is needed regarding intercountry comparisons using input-output data.
As international standards for calculating input-output tables have not yet been
established, inconsistency in the arrangement of data, especially of the service

99

sector, may lead to some estimation errors, making only rough international
comparisons possible.

Since we utilize summarized and aggregated data only, we must

resign ourselves to evaluating column sum or row sum values instead of investigating
the detailed cross-effects that are manifested by the elements of the matrices
themselves.

PART THREE
DUAL ECONOMIC STRUCTURE
CHAPTER 6
THE DUAL STRUCTURE OF THE JAPANESE ECONOMY AND ITS GROWTH PATTERN *

1.

IntJtoducti.oYL

The co-existence of pre-modern and modern economic forms is a "dualistic"


character commonly found in developing countries, not only in production and
distribution methods but also in the mode of life.

In this chapter, the dualistic

character, mainly in the sphere of production methods, will be considered.


As far as the co-existence of pre-modern and modern methods of production is
concerned, Japan and underdeveloped countries in Asia have something in common.

In

Japan, however, pre-modern and modern branches are unified in a national economy,
whereas in Asia's underdeveloped countries they remain marked by the colonial
economy and the mono-culture economy.

In the case of underdeveloped countries, the

term "dualistic state" or "dualism," employed by A. O. Hirschman l ) should be used,


as opposed to "dual structure," a concept which applies to Japan.
Unlike industrial development in Western Europe, where modernization delivered
a frontal attack on pre-modern techniques, in Asian countries modernization may keep
them alive for a long period. This particular tendency has been pointed out by
A. O. Hirschman; in contradiction to many other authors, he maintains that dualism
is "the reason for dualistic development,,,2) which has some compensating advantages
and which represents in a wayan attempt by an underdeveloped country to make the
best of its resources during a transitional phase.

So long as wage differentials

exist between the modern and pre-modern branches, "pre-modern" industries will
probably have an oppotunity of prolonging their life and thus creating a valuable

* Reprinted from The

Veveto~ng Eeono~e6,

with minor revisions.


1) A. O. Hirschman [21], p. 126.
2) Hirschman [21], pp. 125 ~ 132.

Vol. 2 No.2, June 1964, pp. 147

170,

101

period for transition to "modern" operation. On the other hand, existing pre-modern
enterprises run little danger of being eliminated, since new enterprises usually
emerge in an industrial group which is entirely different from the current native
industry. This "dualistic state" could certainly be retained and utilized, but, as
Hirschman conceded, pre-modern branches will sooner or later succumb to the modern
production methods.

It is, however, characteristic of Japan that not only did the

above-mentioned adaptation in utilizing the dualistic state develop, but the process
of adaptation in organizing complementary relations within the framework of the whole
national economy developed as well.
Differentials in modern and pre-modern branches, which are found in Asian
underdeveloped countries, produce a polarization involving high income and low
income. Moreover, these two branches have a strong tendency to coexist in a
heterogeneous and unincorporative form.

It is a main feature of underdeveloped

countries, considered as a stereotype, that traditional and modern branches coexist


without being mixed, like water and oil.

However, in Japan, modern branches include

pre-modern branches in the working mechanism of the national economy:

large modern

enterprises and traditional medium and small enterprises are in a complementary


relation, the latter depending on the development of the former; on the other hand,
large enterprises use the cheap-labor products of medium and small enterprises and
regard them as a cushion against business fluctuations.
Economtc Gltow:th and Vi66Vle~ in Cap.dat IYLtelUl.dy by Size 06 FiJun

II.

1)

Schema

06

Capital Concen.tJr.a..t.i.on and GIlOw:th

06

EYLteJtplU.6e-6

Let us refer to R. F. Harrod s famous "fundamental equati on of economi c growth,"


I

GC

= 8,3)

and apply it to the problem of the growth of enterprises by size of firm.

As is proved by many statistical analyses, the larger the scale of enterprise, the
higher the capital-output ratio (C), whereas the smaller the scale, the lower the
ratio.

Consequently, even in the case where medium and small enterprises would grow

3) R. F. Harrod, [18], in particular Lec. 3, esp. pp. 77ff.

102

at the same rate as large enterprises, the required capital accumulation ratio (s)
necessary for the realization of the growth rate (G) will be small because of the
low capital-output ratio, while on the other hand, the high capital-output ratio of
large enterprises makes the required capital accumulation ratio extremely large.
This means that large enterprises cannot meet the high required capital accumulation
ratio with retained earnings, even though they have a large amount of owned capital.
As a result, they must depend upon outside capital to a considerable extent.
Such a tendency is of special relevance in an economy where, as in Japan, the
ratio of owned capital to total capital is particularly low compared with the
international level. Moreover, the continued high rate of growth after the war has
strengthened this tendency. Large enterprises must now rely on outside funds to a
greater degree than medium and small enterprises.

Here we find a prime motive for

the close connection between large enterprises and banks, and the consequent
concentration of bank loans and discounts in enterprises of larger scale. However,
there are some factors necessary for the realization of this motive.

It is a fact

that larger enterprises have a strong borrowing power as a result of their privileged
position in regard to capital accessibility. This capital accessibility depends,
after all, upon the power of owned capital in large enterprises. Generally speaking
the greater the owned capital, the better the credit rating, and large enterprises
enjoy a favourable position especially in regard to borrowing long-term funds.
Indeed, statistical data indicate that owned capital of large enterprises is greater
than medium and small enterprises not only in absolute amount, but also in the ratio
of owned-capital to working capital (the debit side in the balance sheet).
Even though large enterprises are thus able to actualize a high required
capital-accumulation ratio, the high capital-output ratio itself constitutes an
unfavourable condition.

It means that capital efficiency or capital productivity is

relatively unfavourable, and also that the depreciation cost and interest charge are
high.

However, these disadvantages are eliminated by a high productivity of labor.


In order to explain high labor productivity in the simplest form, let us

consider a "fundamental equation of productivity,"


equation of economic growth.

yo

= k, corresponding to Harrod's

In this equation, y denotes the amount of output per

103

worker,

the capital-output ratio (the average capital-output ratio) and k the

capital intensity (that is, the amount of capital stock per worker).4)
In order to improve in large enterprises the relationship of the capital-output
ratio (0) and high labor productivity (y) which progresses at a greater rate than in
medium and small enterprises, production methods must be adopted so that capital
intensity (k) more than offsets the progress of (0), as indicated by the fundamental
equation of productivity.

Only then can higher productivity be achieved.

In large

enterprises the adoption of high capital intensity means at the same time a high
accumulation of capital stock; it corresponds to capital concentration on the
financial side.
In sum, a high capital-output ratio in large enterprises has two aspects,
financial and material/technical, and these can be disposed in a schema of capital
concentration in larger enterprises (Figure 1).

Figure 1.

SCHEMA OF CAPITAL CONCENTRATION IN LARGE ENTERPRISES

High growth rate


go

=s
Higher ratio of required
Concentration
- - funds
capita 1 accumul ati on

Fi nanci a1
... aspect
Two aspects
of capital
concentration

High capital-output ratio


yo

=k
h
. tt
Concentration of
Material
Hig er capital ln enSl y - - assets in kind .. aspect

High labor productivity

a log k - b

In regard to finances, the high capital-output ratio means an increase in the

4) If

indicates net output,

labor, and

capital, then yo = f~ = ~ = k

104

ratio of required capital accumulation, and this provokes a concentration of funds.


On the other hand, in regard to the technical aspect of production, the high
capital-output ratio means an increase in capital intensity needed for achieving
higher productivity, which leads to a concentration of capital stock. This double
concentration is made possible because large enterprises enjoy capital accessibility.
Mathematical formulas can be used, but first more elements of analysis are required.

2)

Vi66Vle.n.ti.a..t6 in Wa.gu a.nd CapUa,e. Inte.noUy

Looking at capital concentration from the technical aspect, what is the basic
support of the dual structure of the Japanese economy? Various economic, social,
institutional, and historical factors with different influences are at play.

If one

is omitted, the remaining factors are insufficient to give the full picture.
Although there is a real danger in going to extremes, let us concentrate on two
fundamental factors:

1.

the pressure of excess supply of labor or of potential

unemployment, and 2.

the unequal distribution of capital accessibility.

It is a well-known fact that wage differentials in Japan are greater than in


other countries, and the explanation is mostly given from the viewpoint of the
special character of the labor market. On the supply side of labor, there exists
the pressure of excess supply, causing a search for employment at low wages. On the
demand side, medium and small enterprises plan production with low-wage labor,
whereas large enterprises are able to obtain better labor at wages relatively higher
than the difference in quality. On the labor market itself, labor immobility is
characteristic of large enterprises; they have a seniority wage system with automatic
increases according to service years, on the premise of life-long employment. There
is some turnover from large enterprises to medium and small enterprises, but the
reverse movement is out of the question. Under these circumstances, wages in medium
and small enterprises subjected to the pressure of excess supply of labor are low,
and their employees cannot receive the same wages as in large enterprises, even in
cases of long service. The result is a structure of wage differentials by size of
firm.
This view certainly grasps one of the key points of the problem.

However, it

105
has not taken into consideration the conditions of production which give large
enterprises the capacity to play higher wages, nor does it explain Japan's high rate
of economic growth and the permanence of the dual structure.

Differentials in

productivity and the underlying differentials in the composition of capital


accumulation must be examined. 5)
Table 1 gives a summary of the main indicators; their relationship is corroborated in Figure 2.
The upper part of this figure indicates that by locating capital intensity
(tangible fixed assets per employee) by size of firm on the horizontal axis, and
labor productivity (value added per employee or turnover per employee) on the
vertical axis, correlating points (represented by white or black points on the
curves) can be plotted according to the size of firm.
In enterprises of the smallest size (employing less than 10 persons), value
added productivity is 180,000 and capital intensity 70,000, whilst in the largest
(employing 1,000 or more persons), value added productivity is 900,000 to 1,000,000
and capital intensity 600,000 to 700,000.

The difference in capital intensity is

9 to 10 times, causing a difference in productivity of 5 to 6 times.

Between these

extremes, enterprises of the sizes 2,3,4, ... in order of capital intensity form a
convex curve of productivity moving from the lower left to the upper right.

As

shown in the figure, the semi-logarithmic formula y = a log k - b fits the


productivity curve, and the Cobb-Douglas logarithmic formula js also satisfactorily
verified, but the curve fits the semi-logarithmic type better. The realization of
high productivity in large enterprises is assured by a greater increase in capital
intensity than an increase in the capital-output ratio with the enlargement of size
(as illustrated in the lower part of the figure).
Wage differentials are also found in Figure 2.

Locating average annual wages

per employee on the vertical axis, the amount is 100,000 in the smallest class and
300,000 in the largest, the difference being 3 times.

When the average annual

wages are correlated with capital intensity in the direct form (black points on the
5) See Miyohei Shinohara, [45], pp. 103

109.

106

TABLE 1
PRODUCTIVITY, CAPITAL INTENSITY, CAPITAL-OUTPUT RATIO AND
WAGE RATE BY SIZE OF FIRM (1957)

Size by
Number of
Employees

Number
of
Firms

Productivity

Value
added O/L
(thousand
yen per
employee)

CapitalOutput
Capital
Turnover Intensi ty Ratio

Wage Rate

(thousand (thousand Value Turnyen. per yen per added over


employee employee K/O
K/T

(thousand
yen per
employee)

T/L

K/L

fI/L

1-

300,374

186

541

69

0.371

1.128

114

10-

29

77 ,644

289

904

78

0.270 0.086

136

30-

49

13,332

348

1,140

91

0.261

0.080

145

50-

99

8,460

420

1,392

120

0.285 0.086

157

100- 199

3,146

492

1 ,548

166

0.337 0.107

172

200- 299

981

564

1,716

209

0.371

0.122

187

300- 499

645

696

2,088

309

0.445 0.148

205

500- 999

441

780

2,328

408

0.523 0.175

230

9 1,000-1,999

222

922

2,886

589

0.639 0.204

259

10 2,000-4,999

135

1,078

2,872

687

0.669 0.245

301

5,000-9,999

46

866

2,393

558

0.729 0.233

287

28

897

2,643

651

0.727 0.245

329

405,424

516

1,560

289

0.560 0.185

194

11

12 10,000 or more

Total
Source:

Th~

dgta

ar~ tgk~n

from Ministry of International Trade and Industry,

Clm6ha-/agya Saga /ahan Cha,6a. (BIU>,.c. SWtve.y an Me.cUu.m a.nd Smail.

EnteJtpWe.6), 1957.
Note : Manufacturing industry only.
unincorporated firms.

Includes both incorporated and

107
Figure 2.

I-.:j

RELATION BETWEEN CAPITAL INTENSITY AND


PRODUCTIVITY BY SIZE OF FIRM

320

160

C)

>

:;:;
u
-6o

T
L

140

= 2,038.0 log LK
(R = 0.983)

.9

- 2,982

.11

120

100

.=

,..

Q)

'"

SO

,of

:>

Q)

c::

~
'"

3:

--I

20

1
1

lY

,..'"

11

1"12- -

~.---.
_.---

'-~

"S
"S

li
c
n

.12

5 _--.6
a4 ..

'"o

~
""\J

160

,.."C1's

200

__ ,g..-012

40

"S

010

___

~6/---- Q
.. 746 8 log ! - 1,150
L
.
L
2 4 j6 5
(R = 0.987)
3

-:;:

240

'c..

-0

.10 2S0

<T
.....

<
120 .....
~

SO

~
t:-<

J~~----~O
9

40

0.2885 ~ + 115 (R = 0.976)

C>

0r---'1~0---o2~0---o3~0---'4~0~-'5~0~-'6~0~~7rnO~0
~2 ~405 6
Capital Intensity K/L
E-i
10
--2
07
08
........
----__
09
010
:>< 0.2
.3
'-tTi2

2-4

.....o

-(ri1

-+->

'" 0.4
c::
-+->

'"

Co

-+->

.5,

0.6

'"

-+->

'g.

Source:

Note

O.S

Figure 2 is based on the analysis conducted by the author at the


Economic Research Institute, Economic Planning Agency. See K. Miyazawa
and others, "Capital Structure by Fi rm-Si ze," Ec.onomtc. Bu,Ue:Un (i n
English), No.6" edited by the Economic Research Institute, Economic
Planning Agency, Tokyo, 1961. For original data, see Table 1.
The numbers represent the size of the firm based on the number of
employees: the size increases as one goes up from 1 to 12.

108
straight dotted line), a clear linear correlation appears.

Thus, considerable dif-

ferentials in productivity are a cause of wage differentials, and differentials in


capital intensity are a cause of differentials in productivity.

As shown by the

linear correlation formula indicated in the figure, in marginal terms, when the
differential in capital intensity increases by 100,000 as a result of an increase in
scale, wage differentials increase by 28,000 and more.

The strong capital accessi-

bility of large enterprises makes possible the adoption of high capital-intensive


production methods, which in turn make possible high productivity and high-wage
capacity.

On the other hand, medium and small enterprises are forced to adopt labor-

intensive methods with a low technical level (or low capital intensity), due to
their weakness in regard to capital accessibility;

this disadvantage is covered by

low wages, due to the pressure of an excess labor force and potential unemployment.

3)

PeJUnaJ1enc.e 06

:the Vu.ai S:tILu.c.:tuJte

A hypothetical interpretation with regard to the above situation is given in


Figure 3, where differentials in the sphere of production are indicated by the
distance between two dotted curves, namely, the production function in large enterprises (fa) and the production function in small enterprises (fb ).

For simplicity's

sake, differentials in costs are assumed to be represented mainly by wage differentials.


points

Wage rates of large and small enterprises are given on the vertical axis by
a

and b respectively, the difference ab representing the wage differential.

In this case, the production point of maximum profit rate in large enterprises
(obtained by drawing a tangent line from a to the production function f a ) is
in small enterprises
the actual line);

a.

Consequently, the observed productivity curve is

a,

and

(curve of

it is considered to correspond to the actually observed produc-

tivity curve by size of firm illustrated in Figure 2.

In Figure 3, the profit rate

of large and small enterprises (the slope of the tangent line) is supposed to be
equal; but even if differentials in profit rate are supposed, there is no change in
the situation.
If we suppose that, for some reason or another, wage differentials are narrow-

109
Figure 3.

HYPOTHETICAL GRAPH OF THE PERMANENCE OF DIFFERENTIALS

...,

.~

::l

"0

c..

b'
b

O~---------------------------

Capital Intensity

ing, the difference in cost points ab declines to ab'.

In this case, the production

point of small enterprises shifts from S to S', resulting in a decline in the profit rate (the slope of the tangent line).

Consequently, if small enterprises intend,

under the new wage rate, to obtain the same rate of profits as in the past, there
is no other alternative but to adopt high-grade production techniques fi' and raise
the production point to B".

If the wage differential ab' narrows to "zero", and

fi coincides with fa as all enterprises achieve an almost identical level of production, this level would coincide with the production function of the original
J. Robinson type. 6) The reason why J. Robinson succeeded in producing a productivity curve for the economy as a whole is due to the fact that she tacitly presupposed
a homogeneous economy where various differentials are almost negligible.

The actual

conditions in Japan, however, do not warrant such a presupposition.


6) J. Robinson [42J, pp. 81-106.

See also by the same author, [43J, pp. 101 ff.

110
The first fundamental condition to be considered is the existence of differentials in capital accessibility.

Owing to their extreme weakness in capital ac-

cessibility, small enterprises relying on their own strength cannot raise their
capital intensity to a point corresponding to

~"

, nor adopt new techniques (fi).

Under such circumstances, and though many more complex factors will be at work, in
substance there are four possibilities:
profit rate at point
point

~',

(1) to survive, being contented with a low

and generating differentials in profit rate;

(2) remain at

or thereabout by hiring new cheap labor, and failing to narrow the wage dif-

ferentials;

(3) when the above two cases are impossible, small enterprises may

disappear or (4) receive assistance, financial and technical, by subcontracting for


large enterprises, and adopting new techniques

~"

Which of these possibilities has the highest probability? The disappearance of


small enterprises, possibility (3), is of common occurrence; their survival at the
cost of a low profit rate, possibility (1), can also be expected in many cases.

But

in view of the pressure of excess supply of labor, as obtained for a long period
after the war, the survival of small enterprises at a low production point, possibility (2), can be said to have been the most probable case.
ever, points toward possibility (4).

An another trend, how-

In some categories of industry, large enter-

prises undergoing technical innovation tend to develop subcontracting medium and


small enterprises, by giving them assistance in raising funds, providing technical
guidance, lending idle machinery, etc. This results in an improvement in the dual
structure.

However, two qualifications must be added here. Technical innovations

are introduced into the large enterprises themselves which provoke the development
of the subcontracting system. The result is that, according to Figure 3, the shift
to the upper right of the production function in large enterprises f a precedes the
shift to the upper right of fb caused by the modernization of medium and small
enterprises.

Consequently, the trend toward possibility (4) will not necessarily be

sufficient to bridge the gap in the dual structure.


third, and more subcontractors.

Furthermore, there are second,

When large enterprises force a reduction of unit

price, the burden will be passed on further down the line.

111

Thus, even if possibility (4) becomes prevalent, differentials will never be


improved as a whole, as long as unequal distribution of capital accessibility and
pressure of excess supply of labor remain.

The slowing down of the rate of labor

population increase and the increasing trend of demand for labor due to capital
formation are favourable factors, and worthy of attention.

Let us, however, turn to

the problem of capital accessibility.

III.

1)

ViooeJte.n:Ua1l> in Compo.6mon 00 Fu.ncU and InteJte..6t Rate..6

FuncU 00 EnteJtpWe..6 and Cap..ita.l Ac.c.e..6.6ibiUty

In Figure 3, capital accessibility related to the owned capital of the enterprise is presented as one of the shift parameters of the productivity curve fa' fb
Bas~d

on available data in Table 2, the correlation between owned capital Xl and

long-term debt X2 (the amount per person based on the number of workers by size of
firm) is obtained, and amounts to X2

= 0.0667Xll.303.

The elasticity of the long-term

debt to owned capital is 1.30. Elasticity greater than 1.00 means that, in accordance with the expansion of owned capital following the enlargement of size, longterm debts register a greater increase than owned capital.

In other words, when

owned capital rises by one per cent following enlargement of size, long-term debts
increase by 1.3 per cent.

It may be concluded that, for an enterprise the size of

owned capital indicates its degree of capital accessibility,7) and that the competitive position of a firm in the capital market is limited by the amount of owned
capital.
Since the size of owned capital is the fundamental factor for capital accessibility, the larger the amount of owned capital, the more capital-intensive methods
of production are adopted due to the realization of high productivity and inclination towards high capital accumulation.

Such a view is not sufficient to explain

the high capital intensity in large enterprises.

If in Japan the low wage-level is

7) This kind of view is also maintained by M. Kalecki [25], pp. 91-95, and
J. Steindl [49], pp. 40 ff.

112

TABLE 2
DIFFERENTIALS OF CAPITAL INTENSITY, OWNED CAPITAL AND
LONG-TERM BORROWINGS BY SIZE OF FIRM
Size of Firm
Long-term Capital
(total assets) per Employee
( million)
Xl + X2

Owned Capi ta 1
per Employee
Xl

Long-term
Borrowing
per Employee

X2

Capi ta 1 I nten si ty
(tangible fixed
assets per
employee)

0-

70.2

65.6

4.6

58.7

2-

68.3

60.2

8.1

60.2

5-

10

95.7

83.9

11.8

102.0

10-

30

167.7

143.2

24.5

142.6

30-

50

186.0

121.0

65.0

182.2

50-

100

226.9

188.8

38.1

186.0

100-

500

459.7

375,8

85.2

334.8

500- 1,000

640.7

425.8

214.9

533.0

1,000- 5,000

1,014.8

685.7

329.1

739.9

5,000-10,000

1,589.2

1 ,186.6

402.6

1,018.6

1,316.7

1,015.3

301.4

988.9

708.0

529.9

178.1

525.3

10,000Average
Source:
Note

Data are based on Ministry of Finance, Hoji~-KLgyo To~ei Nempo


(YeaJtbOOR 06 CoJtpoJtate EnteJtpWe S-ta.U6:Uc.6), 1957. Manufacturi ng
industry only. Unincorporated firms are not included.
Long-term capital = Owned capital + Long-term borrowings.
Owned capital = Capital + Capital surplus + Earned surplus
(including net profit and loss for the current term).
Long-term borrowings = Corporate debenture + Long-term borrowings
from financial institutions.

113

due to the pressure of excess supply of labor, the adoption of labor-intensive


methods of production would be profitable even to large enterprises, or at least differentials would not be as large as actually experienced.
intensity in large enterprises is high.

Nevertheless, capital

Why? Various factors must be considered.

For instance, Japanese enterprises manifest a strenuous drive to attain the advanced technical levels of developed countries.

However, on the financial side, a

cause is the fact that the price of capital (interest rate) is relatively cheap for
large enterprises.

Adequate data of differentials in interest rates according to the size of


enterprises are not easily available.

In 1958, small enterprises with a capitaliza-

tion of 5 million and less bore an average interest rate of 17 per cent, whilst
large enterprises with a capitalization of 100 million and over used borrowed funds
bearing the relatively low average interest rate of 11 per cent as shown in Table 3.
TABLE 3
AVERAGE LEVEL OF INTEREST RATES ON BORROWINGS BY SIZE OF FIRM
Size classified by paid-in capital -(unit:

Years

per cent)

-2
million

2-5
million

5-10
million

10-50
mill ion

50-100
mill ion

100
mill ion

1956

15.36

14.52

14.28

14.19

13.35

12.24

1957

12.79

15.11

14.40

14.79

12.59

10.23

1958

17.38

17 .80

16.49

13.84

13.62

11.15

Source:

Note:

BaseQ on published data in Ministry of Finance, Hojin-Kigyo Tokei

Nempo (Ye.Mboof< 06 CoJtPoJLa..:te EYLteJtpltMe S~ticA), 1956-58.

Figures are obtained by dividing yearly interest payments by the


outstanding amount of borrowed funds at the end of the year (longand short-term borrowings plus corporate debentures).
The average interest rate in the manufacturing industry = the sum
of interest payment discount + (shoft-term borrowings from financial institutions + long-term borrowings + corporate debentures).

114

Figure 4 gives the cost schedule of raising funds.

Attention is given mainly

to the solid line curve I in the centre. The imputed cost of funds is measured on
the vertical axis. On the horizontal axis, total funds employed are measured, and
are considered to have been put in the order of advantageous sources of investment
funds, namely (a)

internal funds (retained net earnings, depreciation allowances,

etc.), (b) borrowings of various kinds, and (c) equity issues.

Figure 4.

COSTS SCHEDULE FOR INVESTMENT FUNDS

VI

<+-"g
o :::s

of.>LI...
VI

"0

Q)

of.>

:::s

0-

....S

1.

Amount of Investment Funds

Since internal funds may be used freely by the enterprise, their investment

does not incur any cash cost.

However, if these funds are invested outside,

earnings corresponding to the interest rate would be obtained.

In this sense, the

internal investment of these funds means the sacrifice of such external earnings, a
so-called opportunity cost.

If the enterprise has internal funds to the amount of

OA, the accumulation would be made with nearly perfect elasticity to its opportunity

cost.
2.

In case an enterprise is forced to raise external funds, the use of funds

raised through borrowings from financial institutions and the issue of debentures
is accompanied by a cash cost in terms of .interest payments, as well as imputed
costs. This brings about the rise in cost per unit of funds raised up to the point
B.

Real interest rates (cash cost plus imputed cost) rise with the increase in the

amount of debts, because risk premiums are charged.


3. The raising of funds through the issue of stock costs more than other

115
sources of funds, because of the expenses for issuing, the care taken to protect the
market prices of the existing stock, the disadvantages as to taxation, etc.

However,

as long as enterprises accept this high cost, the raising of funds through the issue
of equities would not be so inelastic as in the case of borrowed funds; it may even
be considered as quite elastic. The shape of the curve I on the right side of point
B represents this fact.

This is the graph of the cost schedule for investment funds as shown by J.
Dusenberry.8)

It must, however, be revised in an economy with a dual structure,

because the cost schedule for investment funds shows a great difference according to
sizes of enterprise.
Going back to Figure 4, in medium and small enterprises, owned capital is small
in absolute figures; therefore, as indicated by the broken line lIon the left side,
the gradual increase begins early.

In addition, the weakness of medium and small

enterprises in raising funds results certainly in a rapid advance of imputed costs


of borrowing, and will steepen the slope of the upward curve.

Further, the raising

of funds through the issue of stock and debentures runs into prohibitive
difficulties, due to peculiarities of the Japanese capital market.

Consequently,

the cost schedule for investment funds II ends by being entirely inelastic.
The cost schedule for investment funds of large enterprises is represented by
the broken-line curve IlIon the right side of Figure 4. The amount of internal
funds and the slope of the upward curve are not independent of one another.

It may

be said that the larger the internal funds, the slower the slope of the upward
curve, because, owing to the capital power of large enterprises, borrowed funds can
be raised at lower cost.

Moreover, as large enterprises in Japan have little risk

attached to their borrowing of external funds, it must be considered that this


upward part is actually more elastic than is illustrated, being nearly horizontal.
Further, under the existing circumstances, the cost of expanding net worth through
the issue of stocks being fairly high for large enterprises, it is plausible that
the right-hand part of the curve III, rather than being continuous as in the Figure,
8) J. Dusenberry, [12], pp. 93-99.

116

will in fact be discontinuous and jump to a certain higher level.

At any rate, it

is clear from Figure 4 that the difference between large and small enterprises in
the cost schedule for investment funds causes a marked disparity in the average cost
of raising funds per unit of total capital employed.

Relatively higher interest

rates are paid by small enterprises, and relatively lower rated by large enterprises.
In Japan, an almost institutionalized relationship exists between the different
sizes of enterprise and various types of financial institutions in accordance with
the capital accessibility of the enterprise.

Figure 5 establishes the fact.

This Figure may be called the x-type intersection of borrowed funds of enterprises by size of firm.

Enterprises by size of firm are measured on the

horizontal axis, and ratios of borrowed funds by lenders to total borrowed funds on
the vertical axis. The ratio a (city banks) can be considered as an indicator of
the borrowing power of enterprises in each size group:

it increases regularly as

the size of the enterprise becomes larger; it declines a little at the point of the
largest size, but if borrowed funds from long-term credit banks and trust banks are
added (a'), it shows a smooth upward line.
Ratio b (funds borrowed from customers) can be considered as an indicator
reflecting in part the subcontracting relationship. The curve is high in the case
of medium firms as a reflection of the tendency of medium and small enterprises to
come under the control of large ones.

Ratio c (funds borrowed from money-lenders,

relatives, and acquaintances) and ratio d (funds borrowed from various medium and
small financial institutions) shows a smooth downward line. The weak borrowing
power of medium and small enterprises is manifested by their high degree of dependence on these small financial institutions. The x-type intersection of lines a and
d

is a clear reflection of the dual structure viewed from the financial angle.

In

other words, small enterprises survive by depending for a considerable part of their
funds on money-lenders, customers, relatives and acquaintances, who can be termed
"marginal suppliers of funds." On the other hand, large enterprises subsist in
dependence on the role of the Bank of Japan as a "marginal supplier of funds,"
namely by its advances through the intermediary of city banks to the extent
corresponding to the firms' liquidity position.

Such a situation is practically an

117

Figure 5.

COMPOSITION OF BORROWED FUNDS

%
100

a'

a
50

b ............
00

"''---...

--..... __ .. __

',-

_-

----.. ---.......-........-..-:==~. .-........


.:==.:.-----....-- --.
--_.... ...

1-11-2 2-1 2-2 3

10

11 12

Size of Firm
Source:
Notes:

Same as Figure 2.
1. The size of enterprises on the horizontal axis is measured by the
number of employees, increasing from 1 (l-lO employees) to 12
(lO,OOO and more employees).
.
2. This chart shows the ratio of loans from each lending institution
to total loans:
a - city banks
a'- city banks plus long-term credit banks and trust banks
b - customers
a - b plus money-lenders, relatives, and acquaintances
d - a plus financial institutions for medium and small enterprises.

institutional feature of Japan.

3)

Vi66eJte.rr;tiah, in In:te.Jte,6.t Rate,{) and Unequal. V.i.AtJUbution 06 LoalU

Differentials in interest rate and composition of funds, examined above from


the borrowing side, must now be investigated from the lending side.
Table 4 indicates the average rates of interest by type of financial institutions. The average interest rate of financial institutions for medium and small
enterprises (mutual loans &savings banks, credit associations) are considerably
higher then those of city banks and local banks for loans (loans on deeds, loans on
bills) and discounts (discount of bills): differentials spread from 7.98 per cent

118
p.a. for city banks to 12.47 per cent p.a. for credit associations. These are
averages, of course; rates vary also according to borrowers.
Large enterprises are able to select the lender banks, which is then forced to
put up with interest rates at nearly "competitive prices," whilst for medium and
small enterprises lenders are in a monopolistic position, selecting borrowers and
charging interest at "monopolistic prices."

If it were possible to suppose some

"equilibrium interest rate" which equates with supply and demand of total funds of
all financial institutions, banks as a whole would extend loans to large enterprises
at lower interest rates than the equilibrium rate, and make up the deficits thus
incurred by higher rates on loans to medium and small enterprises.

However, as

regards loans to medium and small enterprises, there is an economic law that the
cost to an individual financial institution of making a loan is higher when the risk
is greater and the amount smaller.
TABLE 4
DIFFERENTIALS IN AVERAGE RATES OF INTEREST BY TYPE OF LENDING INSTITUTIONS (1957)

City banks
Local banks
Long-term credit banks
Trust banks
Mutual loans &savings banks
Credit associations

Loans
%p.a.

Discounts
%p.a.

7.98
8.54
9.41
8.03
10.75
12.47

8.42
9.10
7.87
8.11
10.40
12.47

Source: The Bank of Japan, Hompo Keizai Tokei (Economic


1957.

St~tiC6

06 Japan),

Returning again to Table 4, let us compare the average rate of interest on


loans and the average rate of discount. The textbook argument would be that loans
are longer credit than discounts; therefore rates are to be higher in the former
than in the latter.

However, we find that the average rate of discount is higher

than the average rate of interest on loans in the case of city banks, local banks,
and trust banks.

We may call thi s a "reverse phenomenon ," to be expl ai ned as

119

follows.

The proportion of loans of these banks to large enterprises is high, and

these loans are extended at relatively low interest rates, whilst as to discounts,
the weight of medium and small enterprises is high, and these discounts are made at
relatively high interest rates.

Such a tendency is strongly reflected in the

general average. Thus, large banks are able to extend loans to large enterprises at
relatively low interest rates, because of the high interest rates on loans to medium
and small enterprises.

Furthermore, banks can maintain their liquidity of assets by

offering only short-term loans to medium and small enterprises. On the other hand,
the average rate of interest in long-term credit banks, mutual loans and savings
banks, and credit associations is higher for loans than for discounts, as textbooks
usually maintain. The reason is that customers of these financial institutions are
limited mainly to either larger enterprises (in case of long-term credit banks) and
medium and small enterprises (in case of mutual loans and savings banks and credit
associations).

In this case there is no reverse phenomenon.

Thus, loans of long-term credit banks and others to large enterprises form one
pole, and loans of small financial institutions to medium and small enterprises
another; in the middle, discriminative loans of city banks and local banks to large
and small enterprises produce the above-mentioned reverse phenomenon. This
situation can be ascertained by the difference in the distribution pattern of loans
by various types of financial institutions.
Figure 6 shows the distribution of loans to different sizes of enterprise by
various types of financial institutions, based on data used for Figure 5. The curve
indicated "average" is a cumulative frequency curve for total loans of all financial
institutions, including small financial institutions such as money-lenders and
government financial institutions, which are not listed in the Figure. The uneven
downward development indicates that the distribution of funds as a whole is

unequ~l.

Since in the survey method loans of long-term credit banks and trust banks
cannot be separated, it is regretted that Figure 6 is inconsistent with the data of
Table 4.

However, it appears that the curves of city banks and of local banks are

located between the most uneven curve of long-term credit banks and trust banks
inclining towards large enterprises and the curve of financial institutions for

120

Figure 6.

DEGREE OF CUMULATIVE CONCENTRATION OF LOANS BY SIZE OF FIRM


Financial Institutions for
Small and Medium Enterprises

100r----------------------------+--------~

50

/~

-------'---- Loca 1 Banks


'----1-- Average
----+-- City Banks

----+_
0~~~~==~

__L__L~_ _L__L~_ _~~_ _L_~


4
5 6
7 8
9 10 11 12

Long-term Credit

&Trust Banks

1-1 1-2 2-1 2-2 3

Size of Firm
Source: Same as Figure 2.
Note
The vertical axis indicates the cumulative frequency curve, and the
horizontal axis the size of enterprises by the number of employees.
medium and small enterprises (mutual loans & saving banks, credit associations,
credit co-operatives, etc.).

The above-stated argument on the "reverse phenomenon"

in differentials in interest rate on loans corresponds perfectly to such a


distribution.

So far, a series of differentials in wages, productivity, capital intensity


and interest rate have been observed on the ground of unequal distribution of
capital accessibility.

Let us now re-examine the facts from the technical side of

the production structure.


The decision concerning the degree of combination of the two factors of
production, capital and labor (capital intensity) depends on their relative price

121
which an individual enterprise faces in the factor markets.

In Figure 3, a

theoretical graph of the permanence of differentials, each enterprise's capital


intensity is a function of factor prices.

Medium and small enterprises choose

labour-intensive methods of production with a low capital intensity, in order to


cope with relatively high interest rates and relatively low wages.

On the other

hand, large enterprises adopt capital-intensive methods of production with a high


rate of capital intensity to cope with relatively low interest rates and ,high wages.
This situation can be analyzed by the method of "iso-quant curve of production."
Let us suppose that the production of enterprises of different sizes is
enlarged to the production level of the largest enterprises (with assets of over
10 billion), and the same magnification applies to capital (tangible fixed assets)
and labor (the number of employees) without change in the initial ratio of factor
combination.

The resulting combination of labor and capital is shown in Figure 7.


Figure 7.

ISO-QUANT CURVE OF PRODUCTION

III
....,

OJ

OJ

~
.J:l

.~

0>

'"

III

'"

III

-c
OJ

<: X
.~
...., 4-

.~

0.

'"

Source:
Note

Labor (member of employees)

Data are based on Ministry of Finance, Hojin-Kigyo TORei Nempo


(Ye.aJ1.bOOR 06 Coltpolta;te EnteJtpwe Sta.-U6tico), 1957. Covering
manufacturing industry only.
Enterprises are classified by asset holdings:
1. less than 2 million
7. 100-500 million
2. 2-5 million
8. 500-1 ,000 million
3. 5-10 million
9. l ,000-5,000 mill ion
4. 10-30 million
10. 5,000-10,000 million
5. 30-50 mi 11 i on
11. over 10,000 million
6. 50-100 million

122
Starting from the most labor-intensive combination 1 (enterprise with less than
2 million assets), various combinations 2, 3, 4 ... are determined up to the
highest capital-intensive combination 11.

If enterprises of different size are

acting under the principle of minimum expenses with a production curve identical to
the observed iso-quant curve, the ratio at the tangent line to the various production points on the curve should indicate relative prices of capital and labor. The
textbook argument would be that the greater the slope of the tangent line, the
smaller the capital expenses (interest rates), and the more advantageous it is to
replace labor by capital, and vice versa.
Figure 7 is, to the end, an

ob~e4ved

Nevertheless, the iso-quant curve in

curve of production; it is not the

~chedute

adopted by enterprises of different size as a basis for their behaviour. The


tangent line ought to apply to the individual schedules, but not to the observed
Consequently, it is meaningless to measure the ratio of relative prices of
labor and capital by size of firm from this observed curve. g) The conclusion is the

curve.

following:

enterprises of different size have different iso-quant curves, as

schedules, and the contact points of these respective iso-quant curves and the price
lines of each size of enterprise generate the observed iso-quant curve shown in
Figure 7. This interpretation is identical to that of the observed productivity
curve illustrated in Figure 2.
Since measurement of the schedule iso-quant curves is difficult, the structure
of the ratio of labor and capital prices by size of firm must be obtained by other
means. Table 5 shows that the differentials in the average rate of interest on
borrowed funds (see Table 3) can be converted into an index number taking the
interest rate of the smallest enterprises as 100. The same can be done for the
differentials in wages. 10 ) The ratio of both indices is given as the difference in
g) Such measurement has been tried: the Douglas type production function was
measured from the basic data of Figure 7, and the condition of minimum expenses
added. Then the values of differentials in interest rates were arrived at by
substituting the values of differentials in wages obtained from data using the
minimization equation. However, the values thus obtained could not be judged
significant compared with the actual ones.
10) The differentials in wages are obrained from the Corporate Enterprise Statistics
cited in Table 3.

123
relative prices.
TABLE 5
DISPARITY OF RELATIVE PRICES OF WAGES AND INTEREST RATES BY SIZE OF FIRM
Size of Finn
(paid-in capital)

Differentials in
Wages

( mill ion)

(l,OOO)

(1)

- 2
2- 5
5- 10
10- 50
50-100
100-

139
172
173
200
231
348

100.0
123.5
124.5
143.5
165.9
249.9

Source: Same as Table 3.


Notes : (1) Index number
(2) Index number
the total of
institutions
(3)

(2)

(1)

Differentials in the
Average Rate of
Interest
(%)
(2)
17.4
17.8
16.5
13.8
13.6
10.2

Differentials in the
Ratio of Wages and
Interest Rate
(3 )

100.0
102.3
94.8
79.3
78.2
64.4

1.000
0.828
0.761
0.553
0.471
0.258

of the average wage per employee.


of the discount rate of interest expenses divided by
long-term and short-term borrowings from financial
plus debentures.

The difference in the ratio of relative prices is evident; the ratio of


interest rates to wages is 0.258 for the largest enterprises with a paid-in capital
of over 100 million, whereas it is 1 for the smallest enterprises with a capital of
less than 2 million.

Considering the difference in relative prices in connection

with Figure 7, it may be supposed that the price line of the ratio of interest rates
and wages (as the reciprocal) grows steeper from small to large enterprises.
Different in this from wage differentials, differentials in interest rates are
also common in other advanced countries, because of the risk connected with the
borrower.

In Japan, however, structural aspects are more important.

A fairly

evident type of combination is found between the different sizes of enterprise and
various types of financial institutions.

Secondly, the interest rate has an

important connection with a special feature of capital concentration in Japan,


namely that a firm depends on outside capital rather than on owned capital as in
other advanced countries. Consequently, differentials in interest rates influence

124
considerably and directly the formation of the dual structure.

The 1.00 to 0.258

disparity in relative prices of interest rates and wages (see Table 5) is fairly
large, and it may be concluded that it contributes to the further widening of
differentials in capital intensity.
The analysis of factor ratios in terms of relative prices is a static endeavor.
There are also dynamic factors causing differentials in prices and in capital
intensity.

2)

S:tJtuc.twz.al Pecu,UaJril,[u 06 CapUal Conc.eVLtJtat{.on in Japan

Returning again to general points of view, one aspect has remained untouched in
the above analysis.

Since capital concentration is a phenomenon common to all

capitalist economies, what are the peculiarities of the Japanese case?


Pressure of excess supply of labor and unequal distribution of capital
accessibility have already been pointed out as the two fundamental supports of
Japan's dual structure; they cannot be separated.
tion in Japan presents three peculiarities:

As a result, capital concentra-

capital concentration in a

capital - short economy, the stimul us of the i nternati ona 1 envi ronment, and the
co-existence of large and petty capital.
1.

That capital concentration is related to the pressure of excess supply of

labor means that capital is short in comparison to labor.

Existing capital is

insufficient in regard to the quantity necessary for the whole economy.

As often

pointed out, shortage of capital has been one of the fundamental factors restricting
the development of the Japanese economy.

Capital shortage means an inadequacy of

enterprises to internally generate capital and a need to raise funds indirectly.


Funds scattered in the private individual sector are collected through banks, postal
savings and other financial institutions, which act as intermediaries between
individuals and enterprises.

Such an "indirect financing method," characteristic of

the Japanese money and credit situation, finds its fundamental cause in the shortage
of capital. Thus, the rationale for financial institutions is established, and an
institutional background is provided for differentials in interest rates related to
credit-standing of the borrower.

125

2.

If low wages are due to the pressure of excess supply of labor, the

adoption of labor-intensive methods of production would be advantageous even to


large enterprises. This is a widely accepted economic principle.

Nonetheless,

differentials in interest rates to the advantage of large enterprises exist, and are
a reason for them to adopt high capital intensity. This fact alone does not explain,
however, the existing marked differentials in capital intensity. Also important is
a strong expansion mood and an active inclination towards investments, which tend to
make up for the shortage of capital and which are stimulated by the international
environment in which the Japanese economy has to catch up rapidly with the technical
level of developed countries. This peculiarity is supported in turn by the combination of large banks and large holdings of capital through loans to related
enterprises; differentials in interest rate advantageous to large enterprises can be
considered as cause or effect.

In an economy short of capital, the greater the

degree of concentration of loans to large enterprises, the wider the differentials


in capital intensity between large and small enterprises. Moreover, large enterprises, in case of emergency, can depend heavily on the Bank of Japan as a
"marginal supplier of funds" through large banks.
3.

But what is the situation in regard to small enterprises? In theory, when

capital concentration proceeds, small capital will disappear in the generality of


cases.

In Japan, however, numerous businesses with small capital continue to exist.

Several factors are at play.


ing,excess labor.

Medium and small enterprises offer a means of absorb-

Large and small enterprises are in a complementary relation:

large ones utilize materials manufactured by cheap labor in smaller enterprises


through the subcontracting system and the formation of related enterprises under
their control, and use medium and small enterprises as a cushion against business
fluctuations.

There is also, as already pointed out, the support "marginal

suppliers of funds" can offer medium and small enterprises. The higher interest
rate, due to the greater risk, is a disadvantage compensated by cheap labor. A
last factor would be the high weight of "trade credit" in the composition of the
total funds of medium and small enterprises, which have such recourse to bills and

126
accounts payab1e. 11 )

V.

Conc1.!L6..(.on

In Japan, differentials in wages and interest rates, manifested by the utili'zation of relatively cheap labor in medium and small enterprises, and relatively low
interest rates for large enterprises have solidified the dual structure of the
Japanese economy. An anticipated slowing-down of the growth in the labor supply,
the i ncreas i ng trE!nd of demand for 1abor due to capi ta 1 formati on, and the promotion
of foreign trade liberalization are putting pressure on the pre-modern sectors.
These facts have started a trend towards the abolition of the dual structure.

It

appears imminent that systematization and reorganization of medium and small enterprises will be pushed.

Success depends greatly upon the adaptation and

reorganization of capital accessibility and of the institutional structure of


capital.
In Asia's underdeveloped countries, differentials in income and wages appear as
an extreme difference between the upper and lower classes, somewhat as the division
between two poles.

In Japan, differentials are not as bipolar but moves from a high

stratum to a middle stratum and from a middle stratum to a low stratum, forming a
"continuous differential structure.,,12) The concept of Japan's "dual structure"
must, therefore, be clearly distinguished from the "dualism" found in Asian
underdeveloped countries.
Limiting ourselves to economic factors, the following points are of importance:
1. Duality may characterize the initial stage of development, and be limited
to the formation and utilization of the dualistic development explained by Hirschman,
namely the polarized co-existence of modern and pre-modern branches. However, after
the "take-off," described by Rostow1 3) and especially in the case of rapid progress,
pre-modern branches are to be combined with modern branches in a complementary
fashion.

This is the case of Japan, where indigenous conditions have also

11) This fact is confirmed by various business statistics.


12) Kazushi Ohkawa, [39], pp. 205-217.
13) W. W. Rostow, [44], pp. 25-48.

127

effectively contributed to the adaptation of such a combination.


2.

Considering the national economy as a whole, concentrated utilization of

scarce capital requires the promotion of capital-intensive and modern methods of


production, while low wages due to an abundant supply of labor favour the existence
of medium and small enterprises.
opment.

But here again all depends on the stage of devel-

That underdeveloped countries in Asia have a relationship between capital

and labor resources different from that of developed countries is clearly reflected
in the fact that modern production methods are shut out from various branches of
commerce and industry.

To invest scarce capital in already existing production

activities is too expensive; for an effective use of capital, hope is placed in


investments in industries which introduce new products.
Quite differently, at a more advanced stage of development, as in Japan, the
introduction of capital-intensive methods is required only for the improved
efficiency of

e~ting

production, and if capital is scarce, its distribution among

enterprises will be unbalanced.


3.
aspects.

In the case of Japan, shortage of capital as compared to labor has two


From the international viewpoint, although capital intensity in large

enterprises is said to be high, the level is still lower than in developed countries.
Therefore, these enterprises use a system whereby they complement a productivity
relatively lower than the international level by low wages in subcontracting medium
and small enterprises.

From the national viewpoint, the scarcity of capital brings

about capital concentration in large enterprises in order to use more efficiently a


scarce resource, and at the same time, it brings about the dependence of medium and
small enterprises on labor-intensive methods of production, as well as an incentive
to be subordinated to large-scale enterprises as subcontractors in order to escape
the instability of management resulting from poor and outmoded methods of production.
In this case, the permanence of pre-modern production branches is fundamentally
different from what happens in underdeveloped countries in Asia.
4.

Until the end of the 1950's (or the beginning of the 1960's) the Japanese

economy was characterized by excess supply of labor, and wages in the pre-modern
small enterprises tended to be given by the low level but with some upward trend and

128

cyclical fluctuations.

As a result of the high rate of economic growth during the

1960's, the demand for labor had an increasing trend due to rapid capital formation
and technological progress, and thus decreasing the wage differentials.

One of the

major effects of these developments is the transition from a labor surplus to a


labor shortage economy, which will stimulate the process of adaptation and abolition
of the dual structure.

REFERENCES
[1] R. Artle, "On Some Methods and Problems in the Study of Metropolitan Economies,"
Pap~ and P40ceeding~ 06 the Regional Science A6~o~on, Vol. vii, 1961.
[2] T. Barna (ed.), The stkuctu4al Inte4dependence 06 the Economy, New York and
Milan, J. Wiley &Sons and A. Giuffre, 1955.
[3] A. S. Bhalla, "Role of the Services Sector in Employment Expansion,"
Int~ationat LaboU4 Rev~ew, Vol. 101, No.5, May 1970.
[4] W. H. Branson, "Comments on K. Miyazawa 'Input-Output Analysis and Interrelational Income Multiplier as Matrix' ," EconometUc Society Meeling~ (mimeo) ,
Washington D.C., Dec. 1967.
[5] H. B. Chenery, "Regional Analysis," in [8], 1953.
[6] H. B. Chenery, "Interregional and International Input-Output Analysis," in [2],
1955.
[7] H. B. Chenery and P. G. Clark,
Sons, 1959.

Int~ndUht4y Economi~,

New York, John Wiley &

[8] H. B. Chenery, P. G. Clark and V. Cao-Pinna (eds.), The stkuctU4e and G40wth 06
the I~n Economy, Rome, U.S. Mutual Security Agency, 1953.
[9] J. S. Chipman, The The04Y 06 Inte4-Sect04al Money
Baltimore, John Hopkins Univ. Press, 1951.
[10] C. Clark, The
3rd ed. 1957.

Con~o~

06 Economic

P40g4~~,

Flo~

and Income F04mation,

London, Macmillan, 1st ed. 1940,

[11] G. Debreu and I. N. Herstein, "Non-Negative Square Matrix,"


21, No.4, Oct. 1953.
[12] J. S. Dusenberry,
1958.

Bu6in.~~ Cyct~

Econom~ca,

Vol.

and Economic G40wth, New York, McGraw Hill,

[13] V. Fuchs, The Se4vice Economy, New York, National Bureau of Economic Research,
1968.
[14] J. K. Galbraith, The New

IndUh~

State, Boston, Houghton Miffien, 1967.

[15] W. Galenson, "Economic Development and the Sectoral Expansion of Employment,"


Inte4national LaboU4 Review, Vol. 87, No.6, June 1963.
[16] A. Ghosh, "Input-Output Analysis with Substantially Independent Groups of
Industries," EconometUca, Vol. 28, No.1, Jan. 1960.
[17] R. M. Goodwin, "The Multiplier as Matrix," Economic JouJr.n.ai., Vol. 59, No. 236,
Dec. 1949.
[18] R. F. Harrod,

Towat~

a Dynamic

Economi~,

London, Macmillan, 1949.

130

[19] D. Hawkins and H. A. Simon, "Some Conditions of Macroeconomic Stability,"


Econometnica, Vol. 17, No. 3-4, July-Oct. 1949.
[20] I. N. Herstein, "Comments on Solow's Structure of Linear Models," Econommca,
Vol. 20, No.4, Oct. 1952.
[21] A. O. Hirschman, The
Press, 1958.

S~ategy

06 Economic Vevelopment, New Haven, Yale Univ.

[22] W. Isard, "Interregional and Regional Input-Output Analysis: A Model of a


Space-Economy," Rev.i..w 06 Economico a.nd stctti6:Uco, Vol. 33, No.4, Nov. 1951.
[23] Japan, Ministry of International Trade and Industry, Inten-<-ndUh~al
06 Japanee Economy, (in Japanese), Tokyo, Oriental Economist, 1957.

AnalY6~

[24] Japan, Ministry of International Trade and Industry, Intenneg.i..onal Analy~~


06 Japanee Economy by the 1960 Intenneg.i..onal Input-Output Table, (in
Japanese), Tokyo, Nihonkeizai Shinbunshia, 1967.
[25] M. Kalecki,

Theo~y

06 Economic Vynam.i..c6, London, G. Allen &Unwin, 1954.

[26] N. Kaldor, "Alternative Theories of Distribution," Rev.i..w 06 Economic


Vol. 21, No.6, 1955-56.
[27J J. M. Keynes, The
Macmillan, 1936.

Gen~ Theo~y

06 Employment,

Int~~t,

StucU~,

and Money, London,

[28J W. W. Leontief, The ~~e 06 Am~can Economy, 1919-39, Cambridge, Harvard


Univ. Press, 1941, 2nd ed., New York, Oxford Univ. Press, 1951.
[29J W. W. Leontief, "Interregional Theory," in [30J, 1953.
[30J W. W. Leontief, Input-Output Economico, New York, Oxford Univ. Press, 1961.
[31J W. W. Leontief et al., stucUe .i..n the
York, Oxford Univ. Press, 1953.

~u~e

06 the Amen-<-can Economy, New

[32J K. Miyazawa, "Foreign Trade Multiplier, Input-Output Analysis and the


Consumption Function," Q!,tMtwy JOMnal 06 Economico, Vol. 74, No.1, Feb.
1960.
[33J K. Miyazawa, "Internal and External Matrix Multipliers in the Input-Output
Model ," Hilo~ubMhl JOMnal 06 Economico, Vol. 7, No.1, June 1966.
[34J K. Miyazawa, "Input-Output Analysis and Interrelational Income Multiplier as
a Matrix," Hilo~ubMhl JOMnal 06 Economico, Vol. 8, No.2, Feb. 1968.
[35J K. Miyazawa, "An Analysis of the Interdependence between Service and GoodsProduci ng Sectors," Econom-ic G~owth Cent~ V~CU1>.6-ion Pap~, Yale un-tVeJL6,uy,
No. 92, August 1970, and Hilo~ubMhl JOMnal 06 EconomiC.6, Vol. 12, No.1,
June 1971.

131

[36J K. Miyazawa and S. Masegi, "Interi ndustry Ana lysi s and the Structure of
I ncome-Di stri buti on," Mebl.Oeconom-i.ca, Vol. 15, Fas. 2, 3, Agos to-Di cembre,
1963.
[37J O. Morgenstern (ed.), Econom-tc. Activ~y AnalY4i4, New York and London, John
Wiley &Sons and Chapman &Hall, 1954.
[38J L. N. Moses, "The Stability of Interregional Trading Patterns and Input-Output
Analysis ," AmeiUc.a.n Econom-i.c Review, Vol. 45, No.5, Dec. 1955.
[39J K. Ohkawa, "The Differential Employment Structure in Japan," The Annw 06
the H-Ltot4uba.4hi Academy, Vol. 9, No.2, April, 1959.
[40J K. Ohkawa,

Vi66~enta.t sttuc.t~e

and

Ag~c.uttu4e,

Tokyo, Kinokuniya, 1972.

[41] P. N. Rasmussen, Stu.cLi.u in Ivtt~-Sec.toltal Re1.a.t..i.on4, Kljlbenhavn


Einar Harcks Forlag &North-Holland, 1956.

&

Amsterdam,

[42] J. Robinson, "The Production Function and the Theory of Capital," Review 06
Econom-i.c Stu.cLi.u, Vol. 21, No.2, 1953-54.
[43J J. Robinson, The Accumulation 06 Capital, London, Macmillan, 1956.
[44] W. W. Rostow, "The Take-Off into Self-Sustained Growth," Econom-i.c
Vol. 66, No. 261, March, 1956.

Jo~n.ai..,

[45J M. Shinohara, Gltowth and Cyc.lu in the Japanue Economy, Tokyo, Kinokuniya,
1962.
[46J M. Shinohara, Indu4tnia.t

stJtuc.t~e

(in Japanese), Tokyo, Chikuma-Shoboh, 1966.

[47J R. M. Solow, "On the Structure of Linear Models," Economebtic.a., Vol. 20, No.1,
Jan. 1952.
[48J R. M. Solow, "The New Industrial State or Son of Affluence," The PubUc
InteJtut, No.9, Fall 1967.
[49J J. Steindl, M~ and Stagnation in. Amwc.a.n.
Blackwell, 1952.

Ca.p~m,

Oxford, Basil

[50J R. Stone, "Linear Expenditure Systems and Demand Analysis," Econom-i.c


Vol. 64, No. 255, Sept. 1954.

Jo~n.ai..,

[51] R. Stone, Input-Output a.nd National Ac.c.ount4, Paris, OECD, 1961.


[52] R. Stone and A. Brown, A Computable Model 06 Econom-i.c Gltowth, A PltogJta.mme 60lt
Gltowth, 1, Department of Applied Economics, University of Cambridge, 1962.
[53J F. V. Waugh, "Inversion of the Leontief Matrix by Power Series," Economebtica,
Vol. 18, No.2, April 1950.
[54J M. A. Woodbury, "Properties of Leontief-type Input-Output Matrices," in [36],
1954.

133

INDEX
Advertising, effects of
Ara, K. i i
Artie, R. 75, 129

Employment:
--- expansion in service sector
79, 82
--- requirements 55
Enlarged inverse matrix multipl ier
5, 50
Expenditure of national income 2
External funds 114
External multipl ier 61, 88
--- of regions 66
--- of service sector 91, 98
local ized --- 72

80

Barna, 1. 129
Bhalla, A. S. 76, 79, 129
Borrowed funds, structure of
Branson, W. H. 10, 129
Brown, A. I3 I

116

Cao-Pinna, V. 129
Capital:
--- accessibi I ity I I I
--- accumulation 102
concentration 103, 124
--- intensity 105, 120
--- requirements 55
Capital-labor ratio ~ee Capital
intensity
Capital-output ratio 102
Chenery, H. B. 37, 65, 74 129
Chipman, J. S. 129
'
Circular flow of intermediate
products 44
Clark, C. 79, 129
Clark, P. G. 74, 129
Cobb-Douglas formula 105
Consumption I
--- function 4, 50
--- multipl ier 32
Convergence conditions 18
Cost-push, of service-prices 91
Cost schedule of investment funds

Factor proportion 120


Final demand 2, 29
Financial institutions 116 119
Foreign trade multipl ier: '
modified --- 46
traditional --- 44
Fuchs, V. 80, 129
Funds of enterprises I I I
Galbraith, J. K. 80, 129
Galenson, W. 80, 129
Ghosh, A. 129
Goodwin, R. M. 129
Growth:
--- of enterprises 101
--- of service sector 79
fundamental equation of
114

Debreu, G. 129
Decomposabi I ity 17, 19
Dependence effects 80
Differentials 101
in capital intensity 105, 121
in funds 113, 119
in interest rates 113 117, 123
in productivity 105 '
in wages 104, 120
Direct and indirect requirements 37,55
Distribution of income:
general ized pattern of --- 4
regional --- 29
Distribution of loans 119
Dual structure 108, 126
Dua Ii sm 100
Dusenberry, J. S. 115, 129
Economic growth

~ee

Growth

101

Harrod, R. F. 101, 129


Hawkins, D. 130
Herste in, I. N. 129, 130
Hirschman, A. O. 100, 130
Import:
--- function 47
--- requirements 55
Income distribution ~ee Distribution
of income
Income formation, fundamental
equation of --- 9
Income multipl ier:
interrelational --- 6, 26
macrocosmic --- 2 II
multi-sector --regional --- 29, 36
Income originating (or recievedl
basis 81
Indirect financing method 124
Input-output systems 4, 48
--- and consumption function I, 43

134

--- and price formation 92


general fzation of --- 3
partition off --- 60
Inside propagation ratio:
--- of goods-producing sectors 86
--- of interregional model 66
Interest rate:
different i a lsi n
113, 117, 123
rea I --- 114
Inter-income-group coefficients 8
Interindustry multipl ier ~ee Inverse
matrix multipl ier
Internal funds 114
Internal multipl ier 60, 88
--- of regions 66
--- of service sector 91, 98
re-partitioned --- 71
Interregional model 23
of income formation 29
--- of output determination 37
--- repercussion model 65
I nterre Iat i ona I income mu Itip I i er
6, 26
Inverse matrix:
enlarged --- 5, 50
existence of --- 14
Leontief --- 4, 62
partitioned --- 60
Investment:
--- funds I 14
--- multipl ier 32
I sa rd, W. 65, I 30
Iso-quant of production 121
Kaldor, N. II, 22, 130
Ka I ec k i, M. 2, 2 I, I I I, I 30
Kalecki multipl ier 3, 12
Keynes, J. M. 130
Keynesian multipl ier 2, I I, 56
Labor:
--- market 104
--- requirements ~ee Employment
requirements
excess supp Iy of
I 10, 125
Large enterprises 102, 115
Leakage 32, 36, 45, 68
Leont i ef, W. W. I, 22, 65, 130
Leontief multipl ier 2, 56, 60
Leontief-type matrices 15
Loans:
concentration of --- 120
distribution of --- 119
Local ized partitioned matrix
mu Itip I i er 71
Long-term debt I I I
Marginal suppl iers of funds 116
Marx, K. 76
Marxian approach, on national
products 76

Masegi,S. 1,43,131
Material-coefficients 47
Matrix multipl ier:
--- of income formation 9
--- of output determination ~ee
Inverse matrix
Meaningful solutions, existence of 14
Medium and sma I I enterprises 101, 110,
116
Miyazawa, K. I, 13, 75, 130, 131
Morgenstern, O. 131
Moses, L. N. 37, 65, I31
MPS <material product system) 77
Multi-sector income multipl ier 9
National economic accounting
systems 76
Non-negative matrix 15
Ohkawa, K. 126, 131
Open economy:
--- and input-output model 43
fundamental equation of --- 47
Output multipl ier:
enlarged --- 5
macrocosmic --- 2
partitioned --- 62
regional --- 37
standard --- 4
Owned capital 102, III
Partition off matrices 60, 71
Patrick, H. T. 76
Permutation 18
Pre-modern industries 100
Price formation 92
Production function 108, 122
Productive labor 76
Productivity:
curve 108
--- differentials 105
--- of service 80
fundamental equation of
102
Propagation process:
intersectoral --- 61, 72
structure of --- 12, 46, 56, 78
Propensity to consume domestic
goods 46
Rasmussen, P. N. 131
Regional concentration of incomes 28
Regional income multipl ier 23
Regional model ~ee Interregional model
Relative prices of factors 122
Repercussion process ~ee Propagation
process
Rob i nson, J. 109, 131
Rostow, W. W. I26, I31
Sato, R.

ii

135

Sectoral multipl ier ~ee Inverse matrix


Self-sufficient ratio 32
--- of materials 45
Service:
--- activities 76
expansion multipl ier of --- 78, 81
income-elastisity of demand for
78
interdependence --- and goodsproducing sectors 85
Shinohara, M. 82, 105, 131
Shionoya, Y. 76
Simon, H. A. 130
Simultaneous permutation of matrix 18
Smal I enterprises 101, 116, 125
Smith, A. 76
SNA (system of national accounts) 76
Solow, R. M. 16, 80, 131
Solow's theorem 16
Solutions, existence of ~ee Meaningful
sol utions
Ste i nd I, J. I I I, 131
Stone, R. 10,131
Subcontracting 110, 125
Subjoined inverse matrix 5, 51, 55
Sub-matrix-multipl ier 61, 73
--- of goods-producing activity
88, 93
of price formation 92
--- of service activity 88, 93

regional --- 68
Sub-multipl ier process 44
leakage coefficients of --Suzawa, G. iii

45

Take-off 126
Technical innovation 110
Tertiary sectors ~ee Service
Trade activity 95
Traditional sector ~ee Pre-modern
industries
Transfer of income, between goods and
service sectors 81
Transportation 95
Truncated multipl ier 14
Unproductive labor
Value-added:
mu Itip I i er
ratio 4
sector 10

76

51

Wage differentials 104, 120


Watanabe, T. i i
Waugh,' F. V. 131
Woodburry, M. A. 15,131

Vol. 59: J. A. Hanson, Growth in Open Economies. V, 128 pages.


1971.

Vol. 86: Symposium on the Theory of Scheduling and its Applocations. Edited by S. E. Elmaghraby. VIII, 437 pages. 1973.

Vol. 60: H. Hauptmann, Schiltz und Kontrolltheorie in stetigen


dynamischen Wirtschaftsmodellen. V, 104 Seiten. 1971.

Vol. 87: G. F. Newell, Approximate Stochastic Behavior of n-Server


Service Systems wIth Large n. VII, 118 pages. 1973.

Vol. 61 : K. H. F. Meyer, Wartesysteme mit variabler Bearbeitungs


rate. VII, 314 Seiten. 1971.

Vol. 88: H. Steckhan, GOterstriime in Netzen. VII, 134 Se,ten.


1973.

Vol. 62: W. Krelle u. G. Gabisch unter Mitarbeit von J. Burger


meister, Wachstumstheorie. VII, 223 Seiten. 1972.

Vol. 89: J. P. Wallace and A. Sherret, Estimation of Product.


Attributes and Their Importances. V, 94 pages. 1973.

Vol. 63: J. Kohlas, Monte Carlo Simulation im Operations Research. VI, 162 Seiten. 1972.

Vol. 90: J.-F. Richard, Posterior and Predictive DenSIties for


Simultaneous Equatoon Models. VI, 226 pages. 1973.

Vol. 64: P. Gessner u. K. Spremann, Optimierung in Funktionenrilumen. IV, 120 Seiten. 1972.

Vol. 91: Th. Marschak and R. Selten, General Equilibrium with


Price-Making Firms. XI, 246 pages. 1974.

Vol. 65: W. Everling, Exercises in Computer Systems Analysis.


VIII~ 184 pages. 1972.

Vol. 92: E. Dierker, Topological Methods in Walrasian Economics.


IV, 130 pages. 1974.

Vol. 66: F. Bauer, P. Garabedian and D. Korn, Supercritical Wing


Sections. V, 211 pages. 1972.

Vol. 93: 4th IFAC/IFIP International Conference on Digital Computer Applications to Process Control, Part I. Zurich/Switzerland,
March 19-22, 1974. Edited by M. Mansour and W. Schaufelberger.
XVIII,544 pages. 1974.

Vol. 67: I. V. Girsanov, Lectures on Mathematical Theory of


Extremum Problems. V, 136 pages. 1972.
Vol. 68: J. Loeckx, Computability and Decidability. An Introduction
for Students of Computer Science. VI, 76 pages. 1972.
Vol. 69: S. Ashour, Sequencing Theory. V, 133 pages. 1972.
Vol. 70: J. P. Brown, The Economic Effects of Floods. Investigations of a Stochastic Model of Rational Investment. Behavior in the
Face of Floods. V, 87 pages. 1972.
Vol. 71 : R. Henn und O. Opitz, Konsum- und Produktionstheorie II.
V, 134 Seiten. 1972.
Vol. 72: T. P. Bagchi andJ. G. C. Templeton, Numerical Methods in
Markov Chains and Bulk Queues. XI, 89 pages. 1972.
Vol. 73: H. Kiendl, Suboptimale Regier mit abschnittweise linearer
Struktur. VI, 146 Seiten. 1972.
Vol. 74: F. Pokropp, Aggregation von Produktionsfunktionen. VI,
107 Seiten. 1972.
Vol. 75: GI-Gesellschaft fOr Informatik e.V. Berlcht Nr. 3. 1. Fachtagung Ober Programmiersprachen Munchen, 9.-11. Milrz 1971.
Herausgegeben im Auftrag der Gesellschaft fOr Informatik von H.
Langmaack und M. Paul. VII, 280 Seiten. 1972.
Vol. 76: G. Fandel, Optlmale Entscheidung bel mehrfacher Zielsetzung. II, 121 Seiten. 1972.
Vol. 77: A Auslender, Probh'lmes de Minimax via l'Analyse Convexe et les Inegahtes Variationelles: TMorie et Algorithmes. VII,
132 pages. 1972.
Vol. 78: GI-Gesellschaft fOr Informatik e.V. 2. Jahrestagung, Karlsruhe, 2.-4. Oktober 1972. Herausgegeben im Auftrag der Gesellschaft fur Informatik von P. Deussen. XI, 576 Seiten. 1973.

Vol. 94: 4th IFAC/IFIP International Conference on Digital Computer Applications to Process Control, Part II. ZUrich/Switzerland,
March 19-22, 1974. Edited by M. Mansour and W. Schaufelberger.
XVIII, 546 pages. 1974.
Vol. 95: M. Zeleny, Linear MultlobJectlve Programming. X, 220 pages.
1974.
Vol. 96: O. Moeschlin, Zur Theorie von Neumannscher Wachstumsmodelle. XI, 115 Seiten.1974.
Vol. 97: G. Schmidt, Ober die Stabilitatdes einfachen Bedienungskanals. VII, 147 Seiten. 1974.
Vol. 98: Mathematical Methods In Queueing Theory. Proceedings
1973. Edited by A. B. Clarke. VII, 374 pages. 1974.
Vol. 99: Production Theory. Ed,ted by W. Eichhorn, R. Henn,
O. Opitz, and R. W. Shephard. VIII, 386 pages. 1974.
Vol. 100: B. S. Duran and P. L.
VI, 137 pages. 1974.

O~ell,

Cluster AnalysIs. A Survey.

Vol. 101: W. M. Wonham, Linear Multlvariable Control. A Geometric Approach. X, 344 pages. 1974.
Vol. 102: Analyse Convexe et Ses Applications. Comptes Rendus,
Janvier 1974. Edited by J.-P. Aubin. IV. 244 pages. 1974.
Vol. 103: D. E. Boyce, A. Farhl, R. WeIschedel, OptImal Subset
SelectIon. Multiple Re9ression, Interdependence and OptImal
Network Al90rithms. XIII, 187 pages. 1974.
Vol. 104: S. Fujino, A Neo-Keyneslan Theory of InflatIon and
Economic Growth. V, 96 pages. 1974.
Vol. 105: Optimal Control Theory and Its Applir.atlons. Part I. ProceedIngs 1973. Edited by B. J. Kirby. VI, 425 pages. 1974.

Vol. 79: A. Berman, Cones, Matrices and Mathematical Programming. V, 96 pages. 1973.

Vol. 106: OptImal Control Theory and Its Applications. Part II. Proceedings 1973. Edited by B. J. Kirby. VI, 403 pages. 1974.

Vol. 80: International Seminar on Trends in Mathematical Modelling, Venice, 13-18 December 1971. Edited by N. Hawkes. VI,
288 pages. 1973.

Vol. 107: Control Theory, Numerical Methods and Computer


Systems Modeling. InternatIonal Symposium, Rocquencourt, June
17-21,1974. Edited by A. Bensoussan and J. L. L,ons. VIII, 757 pages.
1975.

Vol. 81: Advanced Course on Software Engineering. Ed,ted by


F. L. Bauer. XII, 545 pages. 1973.
Vol. 82: R. Saeks, Resolution Space, Operators and Systems. X,
267 pages. 1973.
Vol. 83: NTG/GI-Gesellschaft fOr Informatik, Nachrichtentechnische Gesellschaft. Fachtagung ,Cognitive Verfahren und Systerne", Hamburg, 11.-13. April 1973. Herausgegeben im Auftrag
der NTG/GI von Th. Einsele, W. Gilol und H.-H. Nagel. VIII, 373
Seiten. 1973.
Vol. 84: A. V. Balakrishnan, Stochastic Differential Systems I.
Filtering and Control. A Function Space Approach. V,.252 pages.
1973.
Vol. 85: T. Page, Economics of Involuntary Transfers: A Unified
Approach to Pollution and Congestion Externalities. XI, 159 pages.
1973.

Vol. 108: F. Bauer et aI., Suparcritical Wing Sectoons II. A Handbook. V, 296 pages. 1975.
Vol. 109: R. von Randow, IntroductIon to the Theory of Matrolds.
IX, 102 pages. 1975.
Vol. 110: C. Strlebel, Optimal Control of Discrete Time Stochastic
Systems. III. 208 pages. 1975.
Vol. 111: Variable Structure Systems WIth Application to Economics
and BIology. Proceedings 1974. Edited by A Rubertl and R. R. Mohler.
VI, 321 pages. 1975.
Vol. 112: J. Wilhlem, ObjectIves and MultI-ObjectIve DecIsIon Making
Under Uncertainty. IV, 111 pages. 1975.
Vol. 113: G. A Aschinger, Stab,lotatsaussagen uber Klassen von
Matrlzen mit verschwondenden Zeolensummen. V, 102 Selten. 1975.
Vol. 114: G. Uebe, Produktionstheoroe. XVII, 301 Selten. 1976.

Vol. 115: Anderson et aI., Foundations of System Theory: Finitary


and Inflnltary Conditions. VII, 93 pages. 1976
Vol. 116: K. Miyazawa, Input-Output AnalysIs and the Structure of
Income Distribution. IX, 135 pages. 1976.

Okonometrie und Unternehmensforschung


Econometrics and Operations Research
Vol. I

Nichtlineare Programmierung. Von H. P. KUnzi und W Kreile unter


Mitwirkung von w. Oettli. Vergriffen

Vol. II

Lmeare Programmierung und Erweiterungen. Von G. B. Dantzlg. Ins


Deutsche Ubertragen und bearbeitet von A. Jaeger. - Mit 103 Abbildungen. XVI, 712 Seiten. 1966. Geb.

Vol. III

Stochastic Processes. By M. Girault. - With 35 figures. XII, 126


pages. 1966. Cloth

Vol. IV

Methoden der Unternehmensforschung 1m Versicherungswesen. Von


K. H. Wolff. - Mit 14 Diagrammen. VIII, 266 Seiten 1966. Geb.

Vol. V

The Theory of Max-Min and its Application to Weapons Allocation


Problems. By John M. Danskin. - With 6 figures. X, 126 pages. 1967.
Cloth

Vol. VI

Entscheidungskriterien bei Risiko. Von H. Schneeweiss. - Mit 35


Abbildungen. XII, 215 Selten. 1967. Geb.

Vol. VII

Boolean Methods in Operations Research and Related Areas. By P.


L. Hammer (lvanescu) and S. Rudeanu. With a preface by R. Bellman. With 25 figures. XVI, 329 pages. 1968. Cloth

Vol. VIII

Strategy for R&D: Studies in the Microeconomics of Development.


By Th. Marschak. Th K. Glennan Jr., and R. Summers. - With 44 figures.
XIV, 330 pages. 1967. Cloth

Vol. IX

Dynamic Programming of Economic Decisions. By M. J. Beckmann. With 9 figures XII, 143 pages. 1968. Cloth

Vol. X

Input-Output-Analyse. Von J. Schumann. - Mit 12 Abbildungen. X,


311 Seiten. 1968. Geb.

Vol. XI

Produktionstheorie. Von W. Wittmann. - Mit 54 Abbildungen. VIII,


177 Seiten. 1968. Geb.

Vol. XII

Sensitlvitatsanalysen und parametrlsche Programmlerung. Von W. Dinkelbach. - Mit 20 Abblldungen. XI, 190 Selten. 1969. Geb.

Vol. XIII

Graphentheoretische Methoden und ihre Anwendungen. Von W.


Knodel. - Mit 24 Abbildungen. VIII, 111 Seiten. 1969. Geb.

Vol. XIV

Praktische Studien zur Unternehmensforschung. Von E. Nievergelt,


O. MUlier, F. E. Schlaepfer und W. H. LandiS. - Mit 82 Abbildungen.
XII, 240 Seiten. 1970. Geb.

Vol. XV

Optimale Reihenfolgen. Von H. MUlier-Merbach. - Mit 45 Abbildungen.


IX, 225 Seiten. 1970. Geb.

Vol. XVI

Preispolitik der Mehrproduktenunternehmung in der statischen Theorie. Von R. Selten. - Mit 20 Abbildungen. VIII, 195 Seiten. 1970. Geb.

Vol. XVII

Information Theory for Systems Engineers. By L. P. Hyvarinen. - With


42 figures. VIII, 197 pages. 1970. Cloth

Vol. XVIII

Unternehmensforschung im Bergbau. Von F. L. W;lke. - Mit 29 Ab'1ildungen. VIII, 150 Seiten. 1972. Geb.

Vol. XIX

Anti-Aquilibrium. Von J. Kornai. - Mit 31 Abbildunge'1. XVI, 382 Selten.


1975. Geb.

Vol. XX

Mathematische Optimierung. Von E. Blum, W. Oettli. IX, 413 Selten.


(davon 75 Seiten Bibliographie). 1975. Geb.