23 Pages
English

The Evolutionary Explanation of Total Factor Productivity Growth : Macro Measurement and Micro Process - article ; n°1 ; vol.80, pg 93-114

-

Gain access to the library to view online
Learn more

Description

Revue d'économie industrielle - Année 1997 - Volume 80 - Numéro 1 - Pages 93-114
This paper presents the lines along which the analysis and explanation of productivity growth at the macro level can be fundamentally renewed. In an evolutionnary perspective, forces acting on individuals change not these individuals but the structure of the population of which they are members. From this point of view, it is diversity which is at the core of the argument: diversity in firm behaviours, particularly of innovative behaviours, togother with selection mechanisms. Productivity change involves two logically distinct innovation processes, creating new or improved technologies, and diffusion processes, spreading these new technologies into the economic system. On the basis of these principles, the paper proposes some simple formalization for indentifying separately innovation and diffusion productivity effects resulting from the diversity of enterprise behaviours.
Cet article propose les voies selon lesquelles il est possible de renouveler l'analyse et l'explication de la croissance de la productivité au niveau global. Dans une perspective évolutionniste, les forces qui affectent les individus changent en réalité non ces individus mais la structure de la population dont ils font partie. De ce point de vue, c'est la diversité qui est au coeur de l'explication: la diversité des comportements des entreprises, en particulier la diversité des comportements d'innovation, et des mécanismes de sélection. La croissance de la productivité repose sur deux types de processus logiquement distincts: l'un d'innovation renouvelant ou améliorant la technologie, l'autre de diffusion, par lequel l'usage de ces technologies plus efficientes se répand à travers le système économique. Sur la base de ces principes, l'article propose une formalisation simple permettant de dégager les effets de productivité résultant de la diversité des comportements des entreprises, ou distinguant les effets d'innovation et les effets de diffusion.
22 pages
Source : Persée ; Ministère de la jeunesse, de l’éducation nationale et de la recherche, Direction de l’enseignement supérieur, Sous-direction des bibliothèques et de la documentation.

Subjects

Informations

Published by
Published 01 January 1997
Reads 21
Language English
Document size 1 MB

J. Stanley Metcalfe
The Evolutionary Explanation of Total Factor Productivity
Growth : Macro Measurement and Micro Process
In: Revue d'économie industrielle. Vol. 80. 2e trimestre 1997. pp. 93-114.
Abstract
This paper presents the lines along which the analysis and explanation of productivity growth at the macro level can be
fundamentally renewed. In an evolutionnary perspective, forces acting on individuals change not these individuals but the
structure of the population of which they are members. From this point of view, it is diversity which is at the core of the argument:
diversity in firm behaviours, particularly of innovative behaviours, togother with selection mechanisms. Productivity change
involves two logically distinct innovation processes, creating new or improved technologies, and diffusion processes, spreading
these new technologies into the economic system. On the basis of these principles, the paper proposes some simple
formalization for indentifying separately innovation and diffusion productivity effects resulting from the diversity of enterprise
behaviours.
Résumé
Cet article propose les voies selon lesquelles il est possible de renouveler l'analyse et l'explication de la croissance de la
productivité au niveau global. Dans une perspective évolutionniste, les forces qui affectent les individus changent en réalité non
ces individus mais la structure de la population dont ils font partie. De ce point de vue, c'est la diversité qui est au coeur de
l'explication: la diversité des comportements des entreprises, en particulier la diversité des comportements d'innovation, et des
mécanismes de sélection. La croissance de la productivité repose sur deux types de processus logiquement distincts: l'un
d'innovation renouvelant ou améliorant la technologie, l'autre de diffusion, par lequel l'usage de ces technologies plus efficientes
se répand à travers le système économique. Sur la base de ces principes, l'article propose une formalisation simple permettant
de dégager les effets de productivité résultant de la diversité des comportements des entreprises, ou distinguant les effets
d'innovation et les effets de diffusion.
Citer ce document / Cite this document :
Metcalfe J. Stanley. The Evolutionary Explanation of Total Factor Productivity Growth : Macro Measurement and Micro Process.
In: Revue d'économie industrielle. Vol. 80. 2e trimestre 1997. pp. 93-114.
doi : 10.3406/rei.1997.1670
http://www.persee.fr/web/revues/home/prescript/article/rei_0154-3229_1997_num_80_1_1670J. Stanley METCALFE (*)
University of Manchester Centre for Research
on Innovation and Competition,
PREST and School of Economic Studies
THE EVOLUTIONARY EXPLANATION
OF TOTAL FACTOR
PRODUCTIVITY GROWTH :
MACRO MEASUREMENT AND MICRO PROCESS
Mots clés : Productivité totale des facteurs, explication évolutionniste, processus micro
économique, innovation, diffusion, coordination.
Key words : Total factor productivity, evolutionary explanation, micro pocess, innovation,
diffusion, coordination.
T
emergence the our central have between Economic a and neoclassical The residual ignorance measures sought title HE PRODUCTIVITY AND to these growth. Growth of EVOLUTIONARY factor. has to MICRO macro perspective of » this two measure taken (Abramovitz, technical The But paper concepts : growth the a measurement issue PROCESS long and Endless may on progress. GROWTH theory establish is time remains growth, 1956). just how EXPLANATION Search to as after ? No alone live Indeed, a as the In appropriately potentially one controversial part : 1945, down for relationship MACRO does doubts the Understanding this an the not residual is uncountable dominant label that constitute have reflected OF MEASUREMENT between as advancing that TOTAL ever. is been an element 'for it in understanding number embarrassment economic is Indeed, 'Technology the « FACTOR knowledge a literature in relationship measure of the since growth studies growtand the on of to is
h process which does not fit with the demand and supply foundations of factor
substitution and diminishing marginal returns. Not surprisingly, a disproportionat
e amount of effort has been devoted to massaging away the residual with ever
more refined measures of output and factor input. The work of Jorgenson and
Griliches (1967) and Jorgenson (1996) is emblematic in this field of study. Of
course, none of this work claims that advances in technology are not important,
(*) I am grateful to the participants of seminars at IDSE, Milano and Max-Planck-Institut zur
Erforschung von Wirtschaftssystemen, Jena for comments on an earlier draft of this paper.
I thank Sharon Boardman and Deborah Woodman for their considerable help in producing
this draft.
REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 93 rather the claim is that the channels through which new knowledge increases out
put per person are the traditional channels of factor accumulation and factor sub
stitution. Improved capital goods and more skillful workers are treated as equival
ent to more constant quality capital goods and workers, all improvements, in
other words, are embodied and factor augmenting. This line of enquiry is not to
everybody's taste and the question of how to measure input and output remains
endlessly controversial.
The alternative to empirical refinement has been conceptual development as
exemplified in the recent re-emergence of endogenous growth theory. It is an
essential feature of traditional neoclassical growth theory that technical pro
gress is the only source of increases in output per worker in a steady-state cha
racterized by given savings behaviour. As Usher put it, no technical progress
means no growth and no growth implies no capital deepening (Usher, 1980 ;
Nelson, 1981) : a view which restores our measure of ignorance with a ven
geance. Another way to put this is to state that the measured shift in the pro
duction function, the residual, does not measure the total contribution which
that shift makes to economic growth. The « shift » at given factor proportions
induces a greater volume of saving per head of population, and of itself additional substitution of capital for labour. The total contribution of
the shift to the growth of income per head is then the residual divided by the
share of labour income in total output. Moreover in situations of steady growth
with a constant capital output ratio this means that all the increase in output is
accounted for by the « shift » in the production function. In this situation, there
is no independent capital deepening. Rymes (1975) has reached similar
conclusions from an entirely different direction but the net result is the same,
the contribution of technical progress to economic growth is measured by the
increase in output per person and is understated by residual measure of total
factor productivity (1).
Endogenous growth theory has provided an alternative way to address these
questions, either by ruling out diminishing returns to capital accumulation or
by providing an economic explanation of the rate of growth of residual pro
ductivity (Romer, 1990 ; Aglion and Howitt, 1992) (2). The essential feature
of this second approach is to make the rate of residual productivity growth a
(1) Partial productivity rules! Not really, the correct measure should be based on the product
ivity of all the primary inputs but land and natural resources have always been mishandl
ed in models of equilibrium growth. The point, of course, is to distinguish primary inputs
from accumulable inputs the quantities of which are endogenously determined by, among
other things, savings behaviour and aggregate efficiency. For interesting discussion on
land and natural resources see Quadrio Curzio and Pellizzari (1996) and Zoboli (1994).
(2) The so-called 'A.K' models follow the fixed marginal returns to capital route. In so-doing
they constitute a rebirth of the Harrodian theory of growth and development against which
neoclassical growth theory was a reaction. For an excellent survey of this branch of endo
genous growth modelling consult Crafts (1995).
94 REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 function of the scale of the economy by assuming, for example, that a fixed
proportion of the growing population is devoted to the advancement of total
factor (residual) productivity. Typically this story is told in the context of a
research production function which produces the residual. However, as Jones
(1995a, 1995b) has pointed out, the evidence of negligible trends in total fac
tor productivity growth does not fit easily with the strong trend increase in the
proportion of resources devoted to scientific and technological activity, in
advanced countries since 1945.
Moreover, those seeking evidence for non-endogenous growth have been
able to reconcile the cross section empirical evidence with the neoclassical
theory simply by broadening the concept of capital accumulation to include
human capital and so increase the share of total income in GDP from a
figure in the region of 0.3 to one in the region of 0.8 (Barro, 1991 ; Barro and
Sala-i-Martin, 1992 ; Mankiw, 1995). However, as Nelson for one has re
peatedly insisted one has an inevitable degree of freedom in distinguishing
« shifts in » from « relocations around » a production function. Either the pro
duction function shifts, endogenously or exogenously does not matter for this
purpose, or the shape of the production function is such that returns to capital
decline much more slowly than conventional evidence on rates of return and
capital income shares indicates. There is no independent way to separate these
rival hypotheses. At the moment the jury is out although the evidence in favour
of shifting production functions appears persuasive (Islam, 1995).
In the light of the above, I propose to follow a different route and ask « How
is the rate of technical progress in the firm translated into increased producti
vity in the use of resources in the economy as a whole ? I shall argue that total
factor productivity growth is composed of a number of elements and that it is
an inherently statistical concept based upon the aggregation of widely different
behaviours in the economy. We might measure it at the macro level but our
understanding must be built up from the micro level and the co-ordination of
micro behaviours. By statistical I do not mean in the sense of being indeter-
ministic but rather in the sense of the arithmetic, abstract properties of a popul
ation distribution of behaviours. In this I am appealing to well established
evolutionary arguments in which the relevant forces acting on individuals
change not these individuals but the structure of the population of which they
are members. Naturally, this relies on the primitive fact of differences in beha
viour and the central importance of innovation, technological, social, organi
zational, managerial, in creating and recreating the differences on which comp
etitive evolutionary processes can work.
The central aspects of this approach involve the abandonment of the production
function and the idea of smooth factor substitution in given technical conditions.
At the micro level of the firm I will impose fixed coefficients conditional upon the
current state of technical and organizational understanding. Every change in the
coefficients involves a change in knowledge and, to the extent that progress is bia
sed, a concomitant in factor proportions. The process of productivity
change consequently involves two logically distinct and basic processes, the crea-
REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2" trimestre 1997 95 tion of new technologies, de novo or as improvements to existing activities, and
the spread of these new technologies into the economic system. The first we call
innovation, the second diffusion. By diffusion I will mean the process of diffe
rential accumulation by which one technology displaces another. I will not distin
guish it from imitation, the copying of what exists, although a more complete
treatment should do so. Before turning to the analysis I list the claims that I make
in favour of following this route :
— it allows a direct appeal to our understanding of the innovation process and
the immense micro variety of innovative behaviour in modern economies. Only
the complete essentialist and her travelling companion, the uniform agent, could
fail to wonder at the historical record of competing innovations occurring in so
many different sectors and not puzzle as to their economic significance ;
— it allows macro consequences to be modelled as emergent phenomena
conditional upon micro economic variety ;
— it allows us to trace the mutual interaction between patterns of innovation
and structural change and the role of investment as carrier of technology in this
process ;
— it emphasises the critical role of market institutions and the efficiency of
markets for products, primary factors and capital in shaping the translation of
new knowledge into enhanced productivity ;
— it points to the importance of innovations in the demand structure as being
of equal importance with innovations in supply ; and,
— it provides an alternative set of channels through which research, educat
ion and skill formation and the associated policies influence economic growt
h. By making firms the focus in what follows this does not deny the import
ant point that they are embedded within a wider matrix of knowledge genera
ting institutions ranging from other firms, independent research establishments
and higher education institutions.
Central to the explanation is the distinction between inter firm differences in
efficiency and inter firm differences in rates of technical progress. Of course, over
the longer term, the latter underpin the former. In so doing I hope to refocus atten
tion on neglected aspects of the theory of productivity growth and to build a brid
ge with developments in the empirical literature. It is an endogenous growth theo
ry in which micro innovations are translated by economic mechanisms into macro
economic consequences (3). None of the phenomena I highlight could arise in a
uniform world in which all firms within a sector are assumed to employ the same
(3) Following Jones (1995b) it might better be termed a theory of semi-endogenous growth.
96 REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 and necessarily enjoy the same rate of technological change. Indeed, technology,
in such a world, there is no need to distinguish firms, we only need to identify
general production opportunities, usually expressed as a production function. I am
happy to abandon the production function, what I cannot abandon is the diversity
of technological behaviour at the level of the firm. By putting firm behaviours at
the centre of the analysis we achieve precisely what is required to understand the
mechanisms of productivity change.
I will explore and build upon the ideas developed by Nelson, in particular
(1989, 1991), to emphasize the idiosyncratic nature of firm's technologies and
rates of technological change. The core of the argument is concerned with the
explanation of productivity change in a world of diversity in firm behaviour :
and no source of diversity is perhaps more potent than that connected with
innovative behaviour. What is missing from the conventional approach, whether
applied at macro or sectoral level, is recognition of the central role of diversit
y. Not all firms are best practice, far from it, and we have to work in terms of
distributions of behaviour within populations of competing firms. Hence what
productivity growth statistics capture is necessarily change in some suitably
defined measure of average practice. Herein lies the opportunity for a deeper
economic explanation of the relation between change of technology and chan
ge of productivity.
It is a natural concomitant of this approach that the firm and its market
context become central to the analysis. Not only which firms innovate but how
important these firms are in terms of their relative outputs, and how their rela
tive outputs change over time ? The localized nature of technical change matt
ers enormously in determining how average practice evolves (Antonelli,
1994). Whenever firms have different rates of technical progress they must
come to have different unit cost structures. This creates the potential for the
competitive process to drive structural change ; that is, to change the relative
importance of the rival producers. Of course the role of structural change as a
determinant of aggregate residual productivity growth has long been recogni
zed. What has not been sufficiently understood is that structural change is not
an independent element in the picture. Rather it follows from the very diversi
ty in technological behaviour which characterizes competition between firms.
Furthermore, it is only when we have technological differences between
firms that we have any scope for spillovers or imitative behaviour or for that
matter co-operative behaviour. There is simply no point in copying or co-oper
ating with rivals when their behaviours are identical to one's own. If there is
no diversity there is simply no point in even thinking about catching up and
falling behind. Indeed it is self evident that convergence of economic beha
viour is premised upon prior difference in economic behaviour. Hence the cent
ral importance of evolutionary analysis in explaining the change in a world of
different but co-ordinated behaviours. Such phenomena cannot of course be
addressed in terms of a theory of uniform agents, or a theory of states of rest.
While we abandon the idea of equilibrium as state of rest we cannot, of cours
e, abandon the idea of co-ordination of activities in market contexts.
REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 97 provides us with a basis for explaining the rate of productivity growtDiversity
h in two distinct and independent ways : as innovation proper at the level of dis
tinctive productive units, firms ; and as the diffusion of the different technolo
gies based upon the competitive, differential growth of the firms with which
they are associated. These themes are at the heart of the modern empirical lit
erature on technical progress. Now there is no point in imagining that we can ever
have a complete explanation of the first of these elements. Innovation is not
reducible to the predictable consequences of intentional action. While we may
enjoy some success in developing an economic explanation of investments in
innovative activity, this is quite different from explaining the specific kinds of
product, process and organizational innovations which arise. There is simply an
irreducible element of unexplained innovative behaviour which we must accept
(Nelson, 1991). However, we can say much more in economic terms about the
contribution of diffusion. It is diffusion which is the crucial process in ensuring
the spread of innovations to an ever widening circle of economic applications.
It is through these processes that the economic consequences of innovations
flow. It is in this sense that the innovations are primitive while the economic
consequences for employment, investment, and economic structure depend on
the secondary processes. Because diffusion is an process it depends
upon the market and institutional contexts in which it occurs. Consequently,
there is a close connection between the decomposition of productivity growth
advocated here and the wider discussion of innovation processes developed in
recent years by scholars with evolutionary and more eclectic perspectives on the
process of technological change (Abramovitz, 1993 ; Nelson, 1981 ; Fagerberg,
1994 ; Dosi, 1988).
The remainder of this paper builds on these general remarks by developing
an innovation cum diffusion based explanation of productivity growth in the
individual sector and then extending this to derive aggregate productivity
growth from the co-ordination of productivity growth in the different sectors.
A FRAMEWORK
To begin we focus attention on a single economic sector in which a population
of firms produces the same commodity. The firms are identical in all but two
respects. They improve their technology at different rates, whether through tech
nological or organization innovation does not matter, and, as a consequence, at
any point in time they operate different techniques. The cost of capital and
labour is the same for each firm but they do not necessarily sell into a perfect
market for their output. No presumption is made that the firms maximise in a
global sense but they are profit seeking. Each firm is represented as a bundle of
decision routines which characterize its behaviour. These routines relate to cost
structure, innovation activity, investment policy and pricing policy. The firm not
the market sets the product price, and pricing and investment policy play a cent
ral role in explaining the rate of productivity growth.
9g REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 Let there be constant returns to scale and let aj and bj be the unit labour and
unit capital requirements in the ith firm. Set the uniform wage at unity, with
the rental cost of a unit of capital being the required rate of return r (inclusive
of depreciation) multiplied by the price of a unit of capacity, pm. Unit cost,
measured in units of labour, is hj = aj + (rpm) bj. Notice that hj is not equal to
the price set by firm i, unless it happens to be a worst practice firm just brea
king even at the margin of survival in the industry. Firms which are infra mar
ginal generate rents, revenues in excess of costs.
Within the population of firms we have a distribution of unit cost, ranging
from worst practice to best practice. The obvious measure of overall efficien
cy in the industry is average practice unit cost, which may differ markedly
from the cost_ limits at each end of the population distribution. Average practi
ce unit cost hs is a weighted average of the individual costsjhe weights being
the share of each firm in the output of the industry, Sj. Thus hs = S S[hy To fur
ther focus on the essential elements let the capital output ratio be the same for
all the firms in the sector, bj = b.
The rate of change in real unit cost in any firm is
A log h{ = r m{ ± log a{ + (I - m{) d_\ogb 1
dt L dt dt \
[1]
dt dt
that is, it is the sum of a weighted average of the change in unit input require
ments and a weighted average of the change in real factor costs. In this express
ion, mj is the share of labour in unit cost and (1-mj) is the corresponding share
of capital charges (4). Notice that these weights are shares in cost not shares in
sales revenue : labour's share in cost is greater than its share in revenue for all
but marginal firms.
The first bracketed term in [1] is what we mean by total factor productivity
growth or the own rate of technical progress for firm i. It is not generally the
same as the total contribution made by technical change to the reduction of real
unit costs in the firm. Whenever efficiency is increasing in the sector produ-
(4) Factor shares do not, in general, correspond to output elasticities either at the level of the
firm or at the level of the sector. As Nelson (1989) points out, an increase in sectoral
employment can correspond to a wide range of output depending on how the
increase in employment is allocated across the firms. Imagine all the labour going to the
worst practice firm or alternatively imagine it is going to the best practice firm and the point
becomes obvious. Of course, given our assumption the elasticity of output with respect to
labour input is necessarily unity for each firm. Notice also that the decomposition in (1)
does not depend upon relative factor prices measuring marginal rates of factor substitution.
REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 99 units of capacity, the corresponding change in pm will also contribute to cing
the reduction of real unit cost reduction in firm i. We leave these complications
aside. We can help the exposition even further, by assuming that all improve
ments are Harrod neutral at the level of the firm, so b is not only the same for
each firm, it is constant over time. Then the own rate of change in total factor
productivity becomes
£_ lot Tx = -mi ^_log a{
dt dt
which will be different for each firm in general.
However, our interest is not simply in the individual firms but in what is hap
pening to total factor productivity in the sector as a whole. The contribution
which a firm makes to average productivity growth depends upon two things,
its own rate of technical progress and the rate at which the output of that firm
is changing relative to the output of its rivals. As the structure of the industry
changes so does the pattern of resource allocation and with it the measure of
average practice. The point is simple : it is not enough to innovate. For a new
technique to have an economic effect, it must be applied to an ever increasing
share of the output of the sector and this requires it to displace the output pro
duced by other firms. Hence the rate of change of average total factor product
ivity is
^s = I dsi h{ + ls{Éb.
dt dt dt
or
i_ log L = y sJh_ r ± log S{ _ 1 log t{ i m
dt ^ hs Idt dt J
or for future reference,
J_\ogA = - r_^logD-_^ log 7] [3]
dt [_dt dt J
The second part of this expression, the rate of change of T, is the average of own
rates of progress and reflects innovative behaviour at the level of individual firms.
The first part, the rate of change of D, is the diffusion effect, the relative growth
or decline in the importance of different technologies as a consequence of comp
etitive selection. Combined together they give the required measure of total fac
tor productivity growth in the sector. At the level of the firm, output per person,
qj = 1/aj increases at the rate d/dt log q\ = - d/dt log aj, so that the share weigh
ted average rate of increase in output per person becomes
üLlog qs =Ä log âs = _L È. log A [4]
dt dt ms dt
1 00 REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2e trimestre 1997 where = as /7TS is the average share of labour income in the total output ms
of the sector.
Let us now explore in turn the diffusion effect and the innovation effect.
THE DIFFUSION EFFECT
One of the central contributions of the evolutionary theory of economic
change is that it provides us with an endogenous theory of changing
structure based upon competition between rival and different firms. A simpli
fied version of this approach is as follows. Firms expand their productive capac
ity by reinvesting profits so that the rate of growth of each firm, gj, is pro
portional to its unit profit margin. Keeping with our assumption that the only
substantive difference between firms is in their process technology and orga
nization, it follows immediately that gj = f [pj-hj], f being the common pro
pensity to accumulate, and pj the price set by the firm. On the demand side
each firm has a customer base and customers switch between firms in accor
dance to the differences in the prices which each firm sets. Following the
method developed by Phelps and Winter (1970) we can write the rate of growth
of demand for each firm as gj = gß + 8 [ps - p¡\ : gp is the rate of of for the population of firms in this sector, ps is the share weighted ave
rage price, and 8, is a coefficient which measures the degree of market imper
fection (8 = oo corresponds to a perfect market in which all firms must set the
same price). Let firms set their prices to balance the rate of growth of demand
with the rate of growth of supply and this leads to a set of prices which co-ordi
nates the market, in that the rate of growth of output is equal to the rate of
growth of demand in the aggregate.
For our purposes what is important is that this provides an explanation of the
changing structure of the industry. Firms with lower unit costs are more profi
table and set lower prices than higher cost firms, therefore, they attract custom
from rivals and expand capacity more quickly from rivals. Consequently their
particular techniques increase in relative importance and play a larger role in
the determination of average efficiency. In fact, under a regime of balanced
prices, the market shares of the different firms change according to the relation
fLlog si - (g¿ - g) = A r ~hs-hi "I
dt L J
where A = f 8 / (f+8) is the coefficient of selection, the index of the velocity
of the competitive process. The value this coefficient takes depends very much
on the market institutional context. Thus, for example, the propensity to accu
mulate will depend not only upon the retention of profits by firms (their pro
pensity to save) it will also depend upon the supply of external finance and the
role of financial market institutions and the availability of different capital in
struments in the form of debt or equity. Similarly, the institutions of informa-
REVUE D'ÉCONOMIE INDUSTRIELLE — n° 80, 2' trimestre 1997 101