innovation, structural change and demand evolution: does demand saturate?
TRANSCRIPT
INNOVATION, STRUCTURAL CHANGE AND DEMAND EVOLUTION: DOES DEMAND SATURATE?*
Pier-Paolo Saviotti, Faculty of Geosciences, Cross-faculty “Institutions” program, sub-theme “innovation”
Utrecht University, Utrecht, The Netherlands and
GREDEG CNRS, 250 Rue Albert Einstein, Sophia Antipolis, FranceE-mail: [email protected]
and
Andreas Pyka, University of Hohenheim, Economics Institute, Wollgrasweg 23, D-70599 Stuttgart,
Germany, E-Mail: [email protected]
Published in : J Evol Econ DOI 10.1007/s00191-015-0428-2
Abstract: In this paper the exploration of the interaction mechanisms between the evolution of demand and the overall process of differentiation of the economic system is pursued in a twofold way: (i) the question is asked whether demand indeed saturates and whether such saturation is required for an ongoing economic development; (ii) the impact of the income distribution on the evolution of demand and thus on economic development is investigated. To study the existence and impact of demand saturation, Engel curves are calculated under different conditions. To study the impact of income distribution, TEVECON distinguishes within the population of the economic system two social classes, which differ in their level and quality of education, and thus for their human capital and wages.
Keywords Demand . Innovation . Saturation . Income distribution . Engel curves
JEL Classification O O3 O15 E20
1. INTRODUCTION
In this paper we study the impact of the saturation of sectoral demand on the overall pattern of
economic development. The importance of the problem arise for us, starting from our initial
observation that economic development, at least since the industrial revolution, and in particular
since the beginning of the 20th century, had been accompanied by a growing differentiation of the
economic system. We are convinced that this differentiation is not an epiphenomenon of economic
development but one of its fundamental features. In particular we start finding a possible
* The research leading to these results has received funding from the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement n° SSH-CT-2010-266959; project PICK-ME.
1
explanation for the relationship between differentiation and development in Pasinetti's (1981,
1993) proposed imbalance between saturating demand and continuously increasing productive
efficiency. Such imbalance and the consequent bottleneck in development could be overcome by
the innovative emergence of new sectors, which would increase the differentiation of the
economic system. We developed our TEVECON model of economic development on the basis of
this intuition (Saviotti, Pyka, 2004) and focus our attention on the role of differentiation in
economic development. Hence demand saturation plays a central role in the construction of our
model. Many modifications enhancing the interactivity of different components of the economic
system were introduced in subsequent versions of TEVECON. In particular, in a recent paper
(Saviotti, Pyka, 2013) we start investigating the co-evolution of innovation and demand, arguing
that innovation could not have had any impact on economic development, unless demand and
markets are created for the types of goods and services that it helped to produce. In that paper we
explore the transition from a state of the economic system, in which most people could only afford
to purchase the basic necessities of life to one in which most people can purchase a much wider
range of goods and services of increasing quality and of intra sector differentiation. In this paper
we pursue the exploration of the mechanisms, whereby the evolution of demand interacts with the
overall process of differentiation of the economic system by doing two things: first, we ask the
question of whether demand indeed saturates and whether such saturation is required for
economic development to continue; second, we study the impact of income distribution on the
evolution of demand and thus on economic development. To study the existence and impact of
demand saturation, we calculate Engel curves under different conditions. To study the impact of
income distribution, we distinguish in this version of TEVECON within the population of our
economic system two social classes, which differ for their level and quality of education, and thus
for their human capital and wages.
In what follows we will place our paper in the context of the existing literature. In the past, most
growth models, whether of the exogenous (Solow, 1956, 1957) or of the endogenous (Romer,
1990; Aghion, Howitt, 1992) vintages, did not take demand into account. It is only recently, that
some growth models started including demand in their analysis (Matsuyama, 2002; Foellmi,
Zweimuller, 2006). These models differ from TEVECON in a number of aspects and we will
comment on these differences later. Furthermore, some recent papers started studying the effect
of product quality on demand. For example, Bils and Klenow (2001) and Manig and Moneta (2014)
distinguished between quality and quantity Engel curves. Economic development in TEVECON
occurs by means of three trajectories, which involve (i) a growing number of sectors, each based
on a different type of output, and (ii) a growing output quality and (iii) intra-sector differentiation.
In comparing our paper to the literature, we describe how we define output quality and how that
2
2. THE EVOLUTION OF CONSUMPTION
2.1 Past and expected trends
One of the most important trends in consumption in developed countries (DCs) during the 20th
century has been the possibility for large sections of their populations to move beyond the
consumption of basic necessities towards a wider range of higher quality and more differentiated
goods and services (Matsuyama, 2002; Foellmi, Zweimuller 2006; Saviotti, Pyka , 2013). Before
that time, in most societies the consumption of the majority of people was limited to basic
necessities of biological origin, including food, clothing and housing. Such pattern of consumption
did not differ substantially from that of other animal species except for the much greater division
of labour through which such biological necessities were supplied in human societies. Although
more sophisticated patterns of consumption had long been present amongst the few rich people,
for most people a more differentiated consumption including goods and services of higher equality
and of greater intra-sector differentiation had to wait until the 20th century, in what we called the
transition from necessities to imaginary worlds (Saviotti, Pyka, 2013). Starting from this transition,
the share of biological necessities in consumption started falling and the share of artificial,
manmade goods and services started rising. Today biological necessities represent no more than
half of overall consumption.
A growing consumption variety has been accompanied by a growing quality and intra-sector
differentiation of all goods and services consumed, including biological necessities. The increasing
quality and differentiation of the food and clothing, that consumers in industrialized countries
consume today is incomparably greater than that consumed in the 19th century. As far as
biological necessities are concerned, the differentiation of their consumption occurred by
introducing more complex tools (Witt, 2001), which could be considered the capital goods of the
household, ranging from beds to storage devices to cooking utensils, giving rise to what we call
'extended necessities'.
What we call imaginary worlds is in fact a heterogeneous category, including objects such as cars,
computers, telephones, leisure activities, education etc. Although such goods and services are not
biological necessities, they become necessary for survival in manmade environments. Not to have
access to internet can be very debilitating today in some advanced societies. Essentially imaginary
worlds correspond to manmade wants generated by the process of economic development. For
example, the growing participation of women in the labour market benefited from the presence of
household capital goods which could speed up domestic work.
The time allocated to leisure activities has generally increased since the middle of the 20th
century. The beginning of paid holidays gave a considerable impulse to consumption and to the
4
economic activities which provide goods and services complementary to leisure activities. Such
activities are unevenly distributed amongst different countries, even at an equivalent level of
economic development. For example, the USA tend to have much shorter holidays than European
countries (Zilibotti, 2010; Stiglitz, 2010). Such trend in consumption was accompanied by a
considerable reduction in working time, which fell from about 70 hour weekly to about 40 hours
weekly during the 20th century (Bouvier, Diallo, 2014; Whaples, 2001).
An even greater reduction in working times had been predicted by Keynes (1928) who expected
that if previous trends had continued most people in DCs would have been able to solve their
economic problem, with which he indicated the consumption of necessities, by working about
fifteen hours per week. In the remaining time they would have been able to acquire the
consumption habits of the middle classes by indulging in 'higher' activities, such as the arts.
Keynes' excessive prediction of the fall in working times, which seems less excessive if we take
into account the length of holidays in some European countries, was due to the underestimation of
the future possibilities of consumption. The large diffusion of consumer durables, of means of
transport and of communication which occurred throughout the 20th century and which
continues in the 21st absorbed more labour time than predicted by Keynes and reduced the
amount of time which could be allocated to leisure activities. Essentially Keynes did not imagine
the differentiation of the economic system and the consequent expansion of consumption which
would occur later.
Summarizing all these trends we can say, that the general development of consumption in
developed countries during the 20th century corresponds to a movement away from biological
necessities and towards manmade goods and services.
As a consequence we can distinguish four categories of consumption:
1. Biological Necessities
2. Extended biological Necessities
3. Manmade structurally induced wants
4. Leisure activities
If we were today to ask the question 'Will the above trends continue in future and affect all the
countries of the world?' we would have to take into account two factors, which were not
considered in any of the previous papers:
First, the previous papers were focused exclusively on DCs. In LDCs the transition from
necessities to imaginary worlds has for long been limited to the richest classes and it has
been largely or completely absent in the general population.
Second, a factor that could severely limit the continuation of the above trends in
5
consumption, not to mention its possible expansion, is the environmental impact of the
corresponding human activities. This constitutes a sudden reminder of the biological roots
of human beings.
In fact, the two above factors seem to be working in mutually contradictory directions. While the
caching up of LDCs, which would imply the adoption of consumption patterns similar to the past
ones of DCs, would be highly socially desirable, it would aggravate the environmental situation of
the earth, unless it were compensated by a sufficient reduction in the environmental impact of the
required technologies and human activities. To achieve an adequate trade-off between growth
prospects and environmental impact is one of the great challenges of this century which is likely to
require important changes in future consumption patterns.
The scope of these of these trends and of the questions they raise is immense and it is quite
unlikely that any individual model can give us answers to all of them. Our TEVECON model is no
exception and we are going to describe its implications and limitations later. We start here by
referring to the way we represent products, services and quality. Products are not purchased for
their material nature but for the services they perform for their users and consumers. We
represent products by means of two sets of characteristics, corresponding to the internal
structure of the product and to the services it performs (Saviotti, Metcalfe, 1984). This
representation can be considered a variant of Lancaster's (1966) characteristics representation.
In turn, the services supplied by a product correspond to consumers' and users' needs and
wants. In this context the difference between material products and 'pure' services is that the
former supply services are embodied in the internal structure while the latter supply services
are of a disembodied form. Furthermore, output quality is measured by the range and by the
level of the services supplied. In terms of this representation, and discarding the possibility of a
return to the consumption of only necessities, the achievement of an adequate trade-off between
growth and sustainability requires the maintenance of the present level of services combined
with a drastic compensating reduction of their environmental impact.
2.2. Innovation, demand and economic development There is now a large consensus about the role which has been played by innovation in the
growth process occurred since the industrial revolution. However, innovation on its own could
never have had an impact on growth, if the goods and services it produced had not been
purchased by consumers and users or, in other words, if there had not been a demand for the
new goods and services. The emergence of such a demand required (i) a disposable income with
which consumers could purchase these goods and services, and (ii) a set of preferences which
made these goods and services desirable.
6
It seems then clear that a theory of demand needs to address the change in the patterns of
consumption which occurred during economic evolution. Although they have not been included
in the textbook theory of demand, some developments in economic theory tried to address the
problem of the change in the content of consumption. A hierarchical theory of wants has been
discussed by several amongst the leading neo-classical economists, such as Walras (1896, 1988),
Jevons (1924), Marshall (1949), but it has received its most explicit and detailed treatment in
the work of Menger (1950), (Saviotti, 2001). Outside economics a hierarchy of needs has been
proposed by Maslow (1943).
2.2.1 Hierarchical theories of wants The terms demand and consumption are not synonyms. Demand can be expected to refer to
wants and needs while consumption refers to the use of goods and services to satisfy wants and
needs. Consumption and its variations in the course of time can be observed while wants and
needs have to be inferred from these observations. A theory of demand and of its evolution
should start from empirical observations of consumption, in order to infer what needs and
wants consumers had at given times and how these needs and wants changed. The theory of
demand which we find in economics textbooks simply says that consumers maximize their
utility starting from a set of preferences which can be assumed as given and constant. Starting
from such a definition the theory of demand can only be a short run one, unless one assumes
that preferences not only are given but have been constant all throughout human history.
Empirical observations show that consumption changed enormously in the course of time by
integrating completely new types of goods and services. The question then arises of whether
consumers could have preferences for objects of consumption of which they knew neither the
existence nor the properties. There is a growing acceptance that preferences can be endogenous
(Aversi et al. 1999; Stiglitz, 2008) and that they can evolve in the course of time. We do not
attempt to give a systematic answer to this question except for saying that consumers cannot be
expected to have precise preferences for objects of consumption unknown to them. At best we
can assume consumers to have broad and imprecise preferences for particular types of
'services', such as transport, communication, leisure and need to learn how new objects of
consumption could supply these services. The learning process is not be limited to consumers
but needs to involve innovators and entrepreneurs. Following Schumpeter (1912, 1934), we
expect entrepreneurs to detect in a fuzzy form emergent wants, to create goods services to
satisfy them and to teach consumers about their use.
Following Witt (2001) we discuss the evolution of consumption as an evolution of needs and
wants. In this sense, necessities are rooted in biological needs, even if the different ways in
which these necessities can be satisfied can partly be due to wants generated by social
interactions. Additionally, imaginary worlds can be either wants which do not correspond to any
7
biological necessity but are induced by the evolution of the socio-economic system or leisure
activities.
This interpretation of the evolution of consumption bears some similarity to a hierarchical
theory of consumption (Menger, 1950) according to which necessities are lower goods and
services and imaginary worlds are higher goods and services. Such a categorization is
incomplete since it neglects the internal quality and differentiation of the output of different
sectors. Even goods and services corresponding to biological necessities are produced with
growing levels of quality and differentiation. Furthermore, according to Witt (2001) the
consumption corresponding to biological necessities can be coupled with tools or services which
provide enhanced consumption opportunities. We can consider these complementary types of
consumption 'extended necessities'. In this respect, the time trends of necessities and of
imaginary worlds are the same. In both cases we can expect to observe within existing sectors
growth in output quality and in intra-sector differentiation. Various sectors could differ for the
balance between efficiency growth, which reduces costs and prices, and quality growth, which
raises costs and prices. Thus, in rich countries, the share of household expenditures allocated to
food or clothing tends to fall with growing income. Likewise, the prices of electronic goods tends
to fall simultaneously with considerable increases in output quality and differentiation.
Additionally, in rich countries the price of housing tends to rise as a share of household
expenditure. Thus, one of the necessities starts being out of reach for some members of society.
The above trends imply that:
(i) Neither necessities nor imaginary worlds can be treated as homogeneous categories.
Different components of each of them can have different time trends giving rise to structural
change in consumption
(ii) Even if a hierarchical scale of consumption could be established for a population of
consumers there would be considerable differences amongst the members of such a population.
Furthermore, we could expect the extent of variation to be greater for imaginary worlds than for
necessities
What we call imaginary worlds include some very different types of consumption, ranging from
leisure activities to health care to various types of consumer durables. While none of them are in
a biological sense necessities, some of them become 'necessary' in given societies. For example,
not to be able to drive a car or to use a computer can be a serious handicap in modern societies
substantially limiting peoples' ability to find meaningful occupations. In a different sense, the
diffusion of consumer durables such as fridges or washing machines is likely to have been at
least partly induced by the growing participation of women in the labour market.
8
The main point that can be made here is, that the evolution of consumption that occurred over
the last century cannot be due to the emergence of new goods and services for which there has
always been a latent demand. On the contrary, such latent demand and a corresponding set of
preferences can at best have existed in a fuzzy form for some of the present types of
consumption. Other goods and services could simply have been unimaginable. In general we can
expect preferences for new or modified goods and services to be formed by a process of mutual
learning between producers and consumers. The whole process could be started by a
Schumpeterian entrepreneur having an intuition about emerging wants and to continue with
producers needing educate consumers, as Schumpeter thought, until the once emerging goods
and services become so mature that a wide awareness of their existence and knowledge of their
properties are present in a population of consumers. Unsurprisingly, with the availability of this
growing variety in the supply of goods and services marketing has become an essential tool in
many sectors. Furthermore, consumers need to have a growing level of human capital to be able
to consume many of the new goods and services (Saviotti, 2001).
In an interesting paper, somewhat different from the work that made him famous, Keynes’
(1928) explored what he thought would have been the evolution of the capitalist economic
system in the advanced countries of the time. He expected that, if the same trends in technical
progress experienced in the previous century were to continue, most people not only would
have become wealthy enough to have solved what he called 'the economic problem' but could
have done so by working a falling number of hours per week, for example fifteen hours. In fact,
what he meant by the term 'the economic problem' was the satisfaction of what we call
necessities. In Keynes' view most people could have spent the rest of their time in 'higher'
occupations. These higher occupations in Keynes' view seemed not to have an economic nature
(Boldrin, Levine, 2010). Keynes' predictions seemed to be based on the assumption that the
economic system had a fixed composition, in the sense of being constituted by the fixed number
of sectors required to produce necessities with the extent of division of labour that was then
current. However, even without fully accepting Pasinetti's (1981, 1993) assumption of complete
saturation in incumbent sectors, which we will see needs to be relaxed, we now know that the
economic system has become far more differentiated than at Keynes' time and that this growing
differentiation provided an important growth mechanism. What we called imaginary worlds
correspond to a growing number of sectors which have gradually reduced the share of
necessities in overall output and demand. In this context the prediction that working time would
have considerably declined is problematic for two reasons:
(i) It is not certain that if working time had fallen as Keynes predicted it would have been
possible to keep incomes per capita constant or rising. Our work tells us (Saviotti, Pyka, 2004,
9
2008) that growth requires the emergence of new sectors, something which Keynes did not take
into account.
(ii) The new sectors which emerged after Keynes' paper provided the new employment
opportunities that were required to allow the continuation of growth and of increasing incomes
per capita.
As a consequence the emergence of new sectors limited the possible fall in working time, a fall
which nevertheless would have been very problematic without such a change in the composition
of the economic system. Thus, one of the reasons for which Keynes' prediction did not come
true, at least for what concerns the balance between work and leisure time, is the growing
differentiation of the economic system and the employment possibilities it created. Other
reasons for which Keynes' predictions were not entirely correct are discussed in the book edited
by Pecchi and Piga (2010).
In this paper we interpret the changing composition of wants and of patterns of consumption in
terms of the co-evolution of demand and innovation. The differentiation of the economic system
is due to the combination of three trajectories, each one representing the long term trend
towards the rise of efficiency, variety and quality (Saviotti, Pyka 2013). These trajectories are
not independent. In particular, according to Pasinetti (1981, 1993) the saturation of sectoral
demand can lead to a potential imbalance between continuously increasing productive efficiency
and saturating demand. Such imbalance, which could entail a serious bottleneck for economic
development, could be overcome by the creation of new sectors based on new activities and
outputs. In turn, the emergence of new sectors would raise the variety of the economic system.
Thus, the saturation of demand could contribute to the differentiation of the economic system. It
is then quite important to establish whether demand does indeed saturate. To test the existence
of demand saturation we calculate Engel curves under a range of circumstances represented by
different values of TEVECON parameters.
2.2.2 The TEVECON model We are going to explore the evolution of consumption in terms of a model in which economic
development is mainly due to the differentiation of the economic system by the emergence of
new sectors, called TEVECON. Such a model is intended to analyze economic development over
the long run, for example over a period such as the one from industrial revolution to the present.
TEVECON is both an endogenous growth model and a model in which the number of sectors is
endogenously variable. The latter feature is specific to TEVECON and we do not know of any
other model which does that. One of the objectives of the creation of TEVECON was the need to
connect the macroeconomic level of aggregation with micro and with intermediate levels of
10
aggregation. As a consequence TEVECON has both intra and inter sector dynamics. A sector is
defined as set of firms producing a unique although differentiated output. Each sector is created
by a Schumpeterian entrepreneur exploiting a radical innovation in the expectation of achieving
a temporary monopoly. If the innovation is successful, there is going to be a bandwagon of
imitators which is going to reduce the degree of temporary monopoly and to increase the
intensity of competition, thus gradually eliminating the main inducement to entry. In this
process the net rate of entry gradually falls to zero and the rate of exit starts dominating, leading
to a growing industrial concentration. In fact, this process is better described as the transition
from an initially innovative sector to the 'circular flow', the set of routines of an economic
system. With the saturation of a sector we indicate the gradual loss of differential opportunities,
which induce entrepreneurs to enter. This dynamics is represented by a system of equations (Eq
1):
(1) N it=k1 ∙FA i
t ∙ AG it−IC i
t−MA it
where:
Nit = number of firms in sector i
k1 = parameter affected by a number of factors, including the costs of creating a firm, the
presence of entrepreneurs in the economic system etc
FAit = Financial availability in sector i
AGit = adjustment gap in sector i
ICti = intensity of competition in sector i
MAti = failures, mergers and acquisitions in sector i
The rate of entry into a sector i is then determined by the 'financial availability' and by the
‘adjustment gap’ in the same sector. Here 'financial availability' does not just reflect the quantity
of money in the economic system but also the presence of a financial community able to
understand the economic potential of an innovation. The meaning of the ‘adjustment gap’
variable can be understood by thinking that when a radical innovation is first launched there are
neither the demand nor the production capacity for the intended output of the sector. At any
time the ‘adjustment gap’ can be calculated as the difference between maximum potential
demand and the instant demand in the sector. Both demand and production capacity are
gradually created during the life cycle of the sector. As a consequence for a given sector the
‘adjustment gap’ measures the distance from the saturation of the corresponding market. The
adjustment gap is included in the entry term because initially it corresponds to the potential
11
market size. The rate of exit from sector i can be expected to increase with increasing ‘intensity
of competition’ in sector i (ICt) and with the ‘rate of failures, mergers and acquisitions’ in the
sector (MAti). In each sector in a wide range of conditions the number of firms describes a life
cycle, increasing first until a maximum and then falling to a lower value, corresponding to an
increasing industrial concentration and to the saturation of the sector.
We describe the state in which sectoral demand is saturated and in which the intensity of
competition reaches the level prevailing in the incumbent sectors, which corresponds to the
Schumpeterian circular flow, as a state of 'saturation'. Saturation thus conceived plays a very
important role in TEVECON because it represents the loss of profit and growth opportunities,
which had initially induced entrepreneurs' innovative entry. It is this situation of falling
opportunities which induces entrepreneurs to look for other innovations which can be used to
construct new niches and markets. The role played by generalized saturation in TEVECON
justifies the attention we pay to demand saturation in this paper. In turn, the creation of new
sectors represents the inter-sector dynamics of TEVECON.
The number of sectors in TEVECON changes endogenously as entrepreneurs, after having
created a new innovative sector and having it seen to mature and to lose its profit and growth
potential, start looking for other innovations which could lead to new potential markets. If the
saturation of existing sectors represents the inducement to the creation of new ones, new
innovations must be available to create other sectors. The pool of knowledge from which these
innovations can be generated derives from two types of search activities: (i) Fundamental search
activities (FSE) create basic knowledge which can be used in all sectors and speeds up the rate of
their creation; and (ii) sectoral search activities (SEi), which create knowledge specific to each
sector, affect output quality and differentiation and, indirectly, sectoral demand. If the pool of
knowledge can give rise to new innovations when older sectors saturate, then new sectors can
be created, leading to increasing output variety. Both FSE and SEi are generated by the allocation
of economic resources: FSE is funded as part of general investment while SE i is funded by firms
in the sector.
Within TEVECON demand coevolves with innovation and with other aspects of the economic
system. For instance, sectoral search activities increase with increasing demand but search
activities themselves indirectly affect demand through their impact on product quality and on
product differentiation (Eqs 2, 3, 4).
(2) SEit=SE0+k4 ∙[1−exp (−k 5∙ Dacc
t )]
Where:
SEit = sectoral search activities in sector i
12
Dacct = demand accumulated at time t
k4 = parameter representing the effect of accumulated demand on search activities, which we
call technological opportunity
k5 = parameter representing the effect of the speed with which demand affects search activities
Eq 2 shows that demand affects search activities, and Eqs 3 and 4 (see below) show that demand
is affected by Yi and YΔ i,, which in turn are affected by search activities. A further co-
evolutionary mechanism, present only in Eq 4, occurs by means of the creation of the disposable
income required to purchase the output of a new sector in addition to those of the preexisting
ones. This new disposable income is created by the joint effect of the growing productive
efficiency of older sectors and of the income created by the emergence of the new sectors.
We hypothesized that a further co-evolutionary mechanism, in which product quality, wages
and education participate in a feedback loop, contributes to the transition from the low quality
(LQ) to the high quality (HQ) goods and services that we called transition from necessities to
imaginary worlds (TFSC, 2013).
The macroeconomic level of aggregation is calculated by aggregating firm and sectoral
outcomes. As an example of TEVECON results sectoral and aggregate employment are shown in
Figs 1 and 2. There we can see that even when employment creation falls at the sectoral level an
appropriate succession of emerging sectors of the right size and at the right times can
compensate for the falling capacity of individual sectors to create employment.
1 65 129193257321385449513577641705769833897961 -
50
100
150
200
250
300
350
t
Li
Fig 1. The evolution of sectoral employment Li
13
1 57 113169225281337393449505561617673729785841897953 -
100
200
300
400
500
600
t
# aggregate employment & employment trend
Fig 2. The evolution of aggregate employment and trend of employment
The consequence of this dynamics is that the composition of the economic system needs to
change in long run economic development.
The previous considerations show that the evolution of demand cannot be analyzed in isolation
from that of the rest of the economic system. The evolution of the ‘composition’ of demand, or of
the distribution of demand amongst the different types of goods and services that people
purchase at different times, becomes a relevant task for a theory of demand. Such a theory needs
to be constructed in the context of the evolution of the whole economic system. The analysis of
demand, and in particular of its possible saturation, will constitute the main part of this paper.
Before passing to its description we need to describe some more aspects of TEVECON.
The pattern of economic development observed since the industrial revolution has been
accompanied by a considerable differentiation of the economic system. Growing evidence for
this differentiation can be found in a number of references (Frenken et al ,2007; Funke,
Ruhwedel, 2001a, 2001b, 2005; Hidalgo et al 2007; Imbs, Wacziarg, 2003). This process of
differentiation consists of: (i) the emergence of new industrial sectors based on new economic
activities and on new types of goods and services; (ii) the increasing quality and differentiation
of the output of incumbent sectors.
The observed economic development process occurred can be explained by the following three
trajectories, or long term trends:
TR1: growing productive efficiency,
TR2: growing output variety
TR3: growing product quality and differentiation.
14
TR1 consists of the reduction in the inputs required to produce one unit of a constant type of
output. TR2 consists of the increase in the number of distinguishable output types that the
economic system can produce. TR3 consists of the increase in sectoral output quality and intra-
sector differentiation.
The previous trajectories are not independent but complementary. None of them alone could
have generated or explained the observed pattern of economic development. An increase in
productive efficiency alone (TR1) would have reduced the amount of labour and of the other
inputs required to produce one unit of output. Without the emergence of new sectors this would
have been likely to give rise to growing unemployment. Both growing output variety (TR2) and
growing product quality and differentiation (TR3) can compensate the potential labour
displacing effect of TR1, but none of them could have emerged without the increasing productive
efficiency of TR1. Such increasing efficiency gave rise to a surplus which could be invested in the
search activities required to generate TR2 and TR3. In summary, the combination of these three
trajectories led to a growing output variety of the economic system and consequently to a
growing variety of consumption possibilities.
Within this growing range of consumption possibilities different goods and services play
different roles. Some goods can be considered necessities in a biological sense while others are
manmade (artificial) types of consumption which in a past paper (Saviotti, Pyka, 2013) we called
imaginary worlds. The term imaginary means that these new goods and services are not
required for our biological survival but are an artificial construct of human societies.
TEVECON is a very 'spacious' model containing many variables and having gone through a large
number of modification since it first creation. A description of its general aspects has been given
so far. More specific aspects relevant to this and to recent papers will be discussed in section 3.
Given the space constraints of a normal research article we cannot give description of the model
here. A synthetic table containing the most important equations of TEVECON can be found in the
Research Gate pages of both Pier Paolo Saviotti and Andreas Pyka†.
Compared to the papers by Matsuyama (2002) and by Foellmi and Zweimuller (2006), which are
the most similar to ours, our paper differs both for a general methodological choice and for some
more detailed research objectives. TEVECON is an evolutionary model which does not assume
either perfect knowledge or optimizing rationality. Our economic agents are adaptive improvers,
the distribution of whom varies in the population of each economic system. Furthermore, in
TEVCON we can detect regions of parameter space in which development occurs in an ordered
† https://www.researchgate.net/profile/Pier-Paolo_Saviotti https://www.researchgate.net/profile/Andreas_Pyka
15
way and regions in which becomes chaotic or collapses (Pyka, Saviotti, 2011). Like
Matsuyama(2002) TEVECON is based on a dynamical systems modelling approach, but ours is
not derived from the application of a preconceived set of equations but is inductively
constructed by assembling sets of relationships between variables mediated by parameters and
by introducing within these relationships further connections which gradually increase the
interactivity of the of the TEVECON model. These features of TEVCON give us both advantages
and disadvantages. On the positive side we can construct a model of very considerable scope and
flexibility. We can experiment with a large number of variables and with many inter-variable
interactions thus achieving levels of complexity which, although small with respect to the real
world, are greater than those of more constrained analytical or orthodox models. In a sense
TEVECON is intermediate between orthodox (OM) and agent based models (ABM). It has a
greater modelling flexibility than the OM although smaller than that of ABM. On the other hand,
TEVECON is less ad hoc, more overarching and coherent than ABM models. The disadvantage of
TEVCON with respect to OM models is the absence of closure conditions, such as general
equilibrium, which can provide internal means of logical testing of theories and models. TEVCON
can only be tested by comparing predicted with empirical results. Given the generality of
TEVRCON this mode of testing cannot mean the identity of numerical value of system properties
but is limited to the verisimilitude of calculated and observed trends.
In terms of detailed research objectives we differ from Matsuyama (2002) for the choice of
learning mechanisms. Whereas he uses learning by doing we rely predominantly on search
activities, a general analogous of R&D introduced by Nelson and Winter (1982). In this sense,
our model is closer to Foellmi and Zweimuller (2006). However, we differ from them for the use
we make of income distribution: They study its effect on growth while we focus on the impact of
income heterogeneity on the saturation of demand.
We now proceed to the core of this paper, that is to the study of demand saturation and of the
role it can play in economic development.
3. DOES DEMAND SATURATE?
Complete demand saturation occurs when demand stops increasing and remains constant after
a given level of income is attained. This is the type of saturation hypothesized by Pasinetti
extrapolating the work of Engel (1857). In previous papers, we accepted Pasinetti's idea that the
bottleneck originated by the imbalance between saturating demand and continuously increasing
productive efficiency could have been overcome by the emergence of new sectors giving rise to
16
growing output variety. In this sense demand saturation was an essential component of the
mechanism by which economic growth could continue in the long run. Here we challenge this
assumption and ask the questions: Does sectoral demand indeed saturate? Is this saturation
required to allow the long run continuation of economic growth? To answer these questions we
compare the cases of (i) saturation, (ii) under-saturation and (iii) oversaturation. (i) occurs
when demand initially rises and then becomes constant as income rises, (ii) occurs when
demand rises at a diminishing rate without ever becoming constant as income rises, (iii) occurs
when demand falls as income rises. Recent empirical work on Engel curves (Chai, Moneta, 2010,
2011, 2012) shows that complete demand saturation rarely occurs and that non-saturation is a
more frequent outcome, even if more complex Engel curves are observed in a number of cases.
Furthermore, we study the way demand saturation is affected by income distribution.
Following our objective of exploring the co-evolution of demand and of other components of the
economic system, we do not just try to calculate Engel curves but wish to explore the way they
are affected by several non demand variables and parameters. We do not conceive saturation as
an inherent property of demand at a sectoral level, but as a property which is itself determining
and determined by other non-directly demand related economic phenomena.
To achieve the objective described in the previous paragraph we compare the Engel curves
calculated using different versions of the TEVECON model and with different parameter values
in order to see (a) if demand saturation occurs and (b) what circumstances and model
parameters affect the occurrence and type of demand saturation. First, we show the Engel
curves for a version of TEVECON which included a limited extent of co-evolution of demand and
innovation; second, we introduce a stronger form of co-evolution of demand and innovation
combined with the income distribution originating from the distribution of competencies within
two social classes, called low (L) and high (H), differing for their levels of education and of
human capital.
In the first version of TEVECON (v1) sectoral demand depended on product price, product
quality and differentiation, but it is not affected by disposable income (Eq. 1). In this demand
function there was co-evolution of demand and of search activities, but demand itself was not
dependent on disposable income. In the new versions of TEVECON (v2) the sectoral demand
function contains a disposable income term. The two demand functions are shown in equations
3 and 4.
(3) Dit=Di
0 ∙Y it ∙∆ Y i
t
pit
17
-
1
2
3
4
5
6
- 0.02 0.04 0.06 0.08 d1 d2 d3
Engel Curves (kw = 0.25)
-
1
2
3
4
5
6
- 0.02 0.04 0.06 0.08 0.10 0.12 d1 d2 d3
Engel Curves (kw = 2.0)
(4) Dit=Di
0 ∙DDisp, it ∙
Y it ∙∆ Y i
t
p it
where
Dit = demand for good or service i at time t;
Di0 = demand for output type i at the time of its introduction;
DDisp,i = the disposable income which can be allocated to output type i at time t ;
Yit = the quality of output i at times t;
YΔ it = the range within which output quality varies, which measures output differentiation;
pit = the price of output i at time t.
The sectoral demand in Equation 4 will be zero unless the term Ddisp,i, representing the
disposable income for sector i, is greater than zero and its level will be affected by the value of
disposable income. Thus, the term Ddisp,i, in equation (4) adds a further mechanism of co-
evolution to that relating demand to search activities. In this enhanced form of co-evolution of
demand and innovation a new sector i will emerge only if a non zero demand for its output exists
because such non zero demand requires the existence of an adequate disposable income. In turn,
demand increases with innovation because the terms Yi, YΔ i and pi are affected by search
activities (SEi) and search activities are affected by demand. The sectoral demand in equation (3)
is not limited by disposable income and we can expect it to differ from that of equation (4) by
overstating the demand level and possibly by overstating the effect of product quality and
differentiation on demand.
18
Figs 3a and 3b. Engel curves calculated for the TEVECON version v1 with different values of the wages parameter kw. In the two figures the different thickness of the curves corresponds to different sectors. Demand on the vertical axis, total income on the horizontal axis.
The Engel curves in Figs 3a-3b were calculated using demand function (3) and varying the
values of kw, where kw affects wage levels at equivalent labour productivity. In each of these
figures, Engel curves are calculated for three subsequent sectors. We can immediately see that
saturation defined as zero rate of growth of consumption is not attained. The apparent satiation
which seems to occur for very long times is an artificial effect due to the fact that we stop the
runs of our model after a given time. The Engel curves we observe are a series of steps each
constituted by a period of growth followed by a period of quasi constant consumption. The
beginning of each new step coincides with the emergence of a new sector. Thus, if the number of
sectors in the economic system remained constant, Engel curves could achieve saturation.
However, the emergence of new sectors induces a shift in the satiation level of sectoral
consumption by means of the rise in income that it causes. In other words, the shape of Engel
curves is affected by both the tendency to satiation which would occur at a constant number of
sectors and by the shift in the level of satiation induced by the emergence of new sectors. Thus,
satiation is not an implausible assumption as long as the number of sectors is constant, but
becomes implausible due to the structural change, which occurs naturally and which we
maintain is a determinant of economic development.
It is to be noticed that the slope of the growing parts of Engel's curves changes from positive to
negative as the wage rate rises (Figs 3a and 3b). Thus, the higher the wage rate, the more likely it
is, that older goods and services become inferior as they mature during their life cycle. The
period in which the slope of Engel curve becomes negative coincides with that in which the slope
of the income becomes negative after having risen above the trend. In other words, when income
after having grown very fast in the early phases of the new sector starts falling, older goods and
services start being perceived as inferior relative to the new ones. We also observe that the
maximum consumption level of the step corresponding to the emergence of the sixth sector (the
last in our simulation) is not affected by the change in wage rates.
Another novelty of this paper relative to our previous ones is the inclusion of income
distribution amongst the variables, which can potentially affect demand and demand saturation.
We expect that different social groups or classes can have not only different incomes and
incomes per head but also different dynamics of their relative incomes, which potentially affect
Engel curves and demand saturation. The newest TEVECON version (v2) uses the demand
function in Eq. (2) and includes two social classes, differing for their levels of education and of
human capital, giving rise to a distribution of competencies and to a corresponding income
19
distribution. The L and H classes can be compared to blue and white collar workers, or to low
and high skill workers respectively. We can expect the relative income of the two classes to be
affected by the initial distribution of population between the two classes, by the relative
allocation of educational resources, and by the relative quality of education.
If we start with a social system in which the population share of class L is represented by the
term Spop, we can then vary (i) the investment in education as a share of total investment, (ii) the
allocation of this investment to the two classes H and L, and (iii) the relative quality of education
for the two classes. To simplify the comparison, we limit our analysis to the variation of
education's share of total investment (IEd). As we can see from Fig. (4), to vary the education
share of total investment from 5% to 20% considerably affects the evolution of population
shares (Spop). In both cases investing in education raises the share of the H class and reduces that
of the L class. Furthermore, the rise in the share of the H class is much faster and more
pronounced for the higher (20%) than for the lower (5%) education share of total investment.
2 258 514 770 1026128215381794 -
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
t
spop
2 258 514 770 1026128215381794 -
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
t
spop
Fig. 4. Evolution of the population shares (spop) of the two classes L (blue) and H (red) when the
education share of total investment is 5% (Left) or 20% (right). The initial share of the
population in class L is 80% in both cases.
The Engel curves for the two levels of education share of total investment are shown in Figs 5
and 6.
20
- 2 4 6 8 10 12 14 16 18 20 -
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Engel Curves for Low SkilledD
Y - 2 4 6 8 10 12 14 16 18
- 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Engel Curves for High SkilledD
Y
Fig. 5. Engel curves for the L (left) and H (right) classes for a 5% education share of total
investment. Demand on the vertical axis, Income on the horizontal axis.
- 2 4 6 8 10 12 14 -
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Engel Curves for Low SkilledD
Y
- 5 10 15 20 25 30 -
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Engel Curves for High Skilled
Y
D
Fig. 6. Engel curves for the L (left) and H (right) classes for a 20% education share of total
investment. Demand on the vertical axis, Income on the horizontal axis.
The shape of the Engel curves differs substantially between the two social classes. Engel curves
for the L class show oversaturation, in the sense that demand for the output of any sector
gradually falls with rising income. On the other hand, Engel curves for the H class show non-
saturation in the sense that the rate of growth of demand, while remaining positive, falls with
income without becoming constant. We can also notice that the demand level, which sectors
tend to achieve as they move towards maturity, is affected by the education share of total
investment. More precisely, when IEd rises the levels of demand increase for the H class, but fall
21
for the L class. In other words, as investment in education and with it the share of the H class in
population rise, people in the L class have to sacrifice part of their consumption in pre-existing
(incumbent) sectors to be able to consume the output of new and emergent sectors. No similar
sacrifice is required of the people of the H class: They are able to add the consumption of the
output of new sectors while increasing their consumption of the output of incumbent sectors.
The Engel curves for the whole population, combining the consumption of the H and L classes,
show almost demand saturation.
These results can be interpreted as follows:
The saturation of demand hypothesized by Pasinetti does not occur in general. In most of the
situations we explored complete saturation does not occur. The Engel curves calculated with the
v1 version of TEVECON seem to be further away from saturation than those calculated with the
v2 version. This can be explained by the fact that the absence of the disposable income term in
the demand equation (3) can be expected to overstate demand levels, thus leading to excessively
large steps due to the emergence of new sectors. On balance, we expect the Engel curves
calculated with the v2 version of TEVECON and including disposable income to be closer to real
ones.
A relatively common behaviour detected so far in our experiments consists of either under-
saturation or oversaturation, where the former is defined as a progressive fall in the rate of
growth of sectoral demand with income as a sector matures, while the latter is defined as a
progressive rise in the rate of growth of sectoral demand with income as a sector matures. In
both cases the rate of growth of demand with income is very high in the initial phases of the
industry life cycle, corresponding to a situation in which the sector is far from saturation, and
starts falling at a time which seems to correspond to the onset of the saturation of the sector.
The shape of Engel curves differs considerably for the two social classes L and H. For example,
under a wide range of circumstances the Engel curves of the L class show oversaturation while
those of the H class show non-saturation. Thus, demand saturation for the whole population of
an economic system could be an artificial result of the aggregation of individual demands, none
of which saturates. The shape of the aggregate Engel curves depends on the relative size of the
two classes, and on their endowment and quality of education. This implies that although higher
levels of education are an advantage for society as a whole, those members who cannot get
access to education become relatively disadvantaged.
The shape of Engel curves is affected by wage levels and by technological opportunity. This is
logical since both of these TEVECON model parameters affect directly or indirectly demand.
22
Furthermore, the very same process of differentiation underlying economic development can
affect Engel curves. If the number of sectors were fixed and if output could somehow continue to
grow, there would be no need for any reduction in expenditures on existing goods and services
to buy emerging ones.
As we have expected, the shape of Engel curves is not an intrinsic and invariable property of
demand but is affected by the interactions of demand with many other components of the
economic system and by many TEVECON model parameters. This is completely natural when
there is co-evolution of many or most components of the economic system. As a consequence the
dynamics of demand, including its possible saturation, cannot be separated from the dynamics of
the whole economic system. In the co-evolution of innovation and demand, saturation can be
both, a cause and an effect of the emergence of new sectors, contributing to their emergence but
being affected by their presence and dynamics later on.
The previous results have considerable implications for processes of economic development.
According to Pasinetti (1981, 1993) the emergence of new sectors can be induced by the
imbalance between continuously growing productive efficiency and saturating demand.
However, complete saturation of demand is not required. An imbalance can occur even if the
rate of growth of demand is inferior to the rate of growth of productive efficiency. In this case we
can expect the intensity of all the inputs required to satisfy demand to fall over time, thus
entailing a gradual fall in the capacity of the incumbent sectors to create employment. This
induces a shift in investment from incumbent mature sectors to emerging ones. Furthermore, we
have to take into account that an economic system in which search activities are carried out all
the times, can create opportunities which could be exploited by entrepreneurs, irrespective of
the above imbalance between productive efficiency and demand. In other words, while in
Pasinetti's ideas the emergence of new sectors is induced by a form of diminishing returns to the
exploitation of incumbent ones, a further mechanism of economic development are driven by
the opportunities presented by search activities. The two mechanisms are described as threat
based or opportunity based. The former bears a considerable resemblance to the mechanism
which according to Nelson and Winter (1982) leads to most types of innovation: The failure to
achieve existing targets based on present routines induces firms to switch to different routines
using the results of search activities.
To this we add that the concept of demand and market saturation is considerably amplified by
the growing product/output quality and differentiation. This introduces a difference between
volume and value saturation. The former would be attained, when the output of a sector -
measured in volume terms - satisfies a demand, which in this respect is constant. The latter is
23
attained, when the quality and the differentiation of the output of a sector can no longer be
increased. Given that the price of a product is likely to be positively related to its quality and that
product quality and differentiation can be expected to grow together, we can expect the average
price to rise and its distribution to become wider within a given sector in the course of time. The
attainment of value saturation then depends on technology, which reduces unit costs and
increase product quality and differentiation, and on disposable income. As a consequence of the
previous considerations a sector, which has attained volume saturation, has not necessarily
attained value saturation. In general, we can expect value saturation to be attained after, and
possibly long after, volume saturation.
The introduction of product quality and differentiation modifies the possibility of compensation
for the imbalance identified by Pasinetti. Whereas in Pasinetti compensation for the imbalance
between productive efficiency and demand comes from the emergence of new sectors, and thus
from growing output variety, in our TEVECON model, compensation comes from both (i) the
emergence of new sectors and from (ii) the growing product quality and differentiation within
each of them. This leads to the possibility of multiple development paths arising from different
combinations of (i) and (ii). Furthermore, the development paths based on different
combinations of (i) and (ii) entail substantial differences in a number of variables such as wages
and the quality of human capital, and in macroeconomic outcomes such as the rates of growth of
income and employment (Saviotti, Pyka, 2013).
4. DEMAND CURVES, OUTPUT QUALITY AND CONSPICUOUS CONSUMPTION
The trend towards rising product quality and differentiation is also likely to affect demand
curves. Such growing quality and differentiation can only occur if consumers have a disposable
income allowing them to purchase the corresponding products. In a previous paper (Saviotti,
Pyka, 2013) we have shown that such disposable income can be created by the combination of
growing productive efficiency in incumbent sectors and by the employment effect accompanying
the emergence of new sectors. We can then expect that as societies become richer, a growing
share of their income will be spent on (i) new products and services and on (ii) higher quality
and more differentiated products and services. As a consequence the distribution of product
quality and prices within the offered range changes with the average (or median) product
moving from the cheapest to higher priced ones. To test this hypothesis we carried out some
experiments in which we changed the extent to which product quality and differentiation are
increased by search activities. In the present version of TEVECON we cannot calculate demand
24
curves since the model, being designed to study long range phenomena, calculates one demand
point at each time. We can calculate quasi demand curves for extended periods of time of
variable length. We do this using different values of the parameters affecting product quality and
product differentiation. The results are shown in Fig 5, where we see that when search activities
lead to high rates of growth of product quality and product differentiation demand curves are
not uniformly downward sloping as price rises, but contain a section in which demand rises as
price rises. A similar behaviour is found for many types of differentiated goods, such as cars,
watches, television sets, washing machines, etc.
The section in which demand and price rise together correspond also to the behaviour observed
by Veblen (1899) (Trigg, 2001) and are called conspicuous consumption, according to which
consumers do not buy what they need in a functional sense but to impress other people and to
establish their social status. While before the industrial revolution this used to be true only for
very rich and powerful people, with the growing wealth generated in industrialized countries
such a form of consumption has become very common. Indeed, the cheapest goods in each
sector are not those which sell the most. Figure 7 shows that the possibility of conspicuous
consumption increases, when innovation generates high rates of growth of product quality and
product differentiation of goods and services. Given the extent to which this phenomenon has
become widespread, it is doubtful whether the term conspicuous consumption should be used
for the whole range in which demand rises with rising price or just for extremely expensive
products. Further experiments are required to clarify this type of behaviour.
- 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
D
P
25
Fig 7. Effect of changing the parameters affecting product quality (k14) and product
differentiation (k15) on the shape of demand curves. Product quality and product differentiation
increase in the order dark blue, light blue, violet, green, red. Demand on the vertical axis, time on
the horizontal axis.
5. SUMMARY AND CONCLUSIONS
In this paper we study the evolution of consumption activities by means of our TEVECON model
of economic development. In previous papers (Saviotti, Pyka, 2012, 2013) we showed that the
range, quality and internal differentiation of goods and services increased enormously since the
industrial revolution and in particular since the beginning of the 20th century. Furthermore, we
had already stressed that innovation alone cannot explain economic development unless
demand co-evolves with it. In this paper we analyze the mechanism by means of which demand
and innovation interact, and in particular the role of demand saturation in inducing the
emergence of new goods and services. We find that complete saturation, in which demand first
rises but then becomes constant as income rises, is a relatively infrequent behaviour. Deviations
from saturation, and consequently the shape of Engel curves, are affected by several factors. For
example, the presence or absence of an income term in the demand function leads to
considerable changes in the shape of Engel curves. Saturation of sectoral demand is never
achieved because the emergence of new sectors keeps shifting the potential saturation level
which is attained, if there was a constant number of sectors. Additionally, when the demand
function contains a disposable income term and there are two social classes differing for their
education levels and competencies, Engel curves show either over-saturation or under-
saturation, where the former leads to a fall in the rate of growth of demand with rising income
and the latter leads to a slowdown in the rate of growth of demand when income rises. We will
need more experiments to explore the potential impact of other variables and parameters on the
shape of Engel curves. However, even from these limited experiments, we can conclude that the
relationship between demand and income and the shape of Engel curves are not intrinsic
properties of demand, but that they depend on the interactions of demand with a number of
other variables of the socio-economic system. This reinforces our conviction that co-evolution is
a fundamental mechanism of economic development, which is essential to understand long run
processes of structural change and transformations.
REFERENCES
26
Aghion P., Howitt P. (1992), A model of growth through creative destruction, Econometrica, Vol. 60, 323-351.
Aversi R, Dosi G, Fagiolo G, Meacci M, Olivetti C (1999) Demand dynamics with socially evolving
Bils M., Klenow P.J.(2001), Quantifying Quality Growth, The American Economic Review, Vol. 91, No. 4, pp. 1006-1030
Boldrin M., Levine D.K., (2010), All the interesting questions, almost all the wrong reasons, in Pecchi L. Piga G. (Eds) Revisiting Keynes: Economic Possibilities for our Grandchildren, Cambridge MA, MIT Press
Bouvier G., Diallo F. (2014), Soixante ans de réduction du temps de travail dans le monde , INSEE http://www.insee.fr/fr/themes/document.asp?ref_id=ip1273
Chai, Andreas, Moneta Alessio (2010), "Retrospectives: Engel Curves," Journal of Economic Perspectives, American Economic Association, vol. 24(1), pages 225-40, Winter.
Chai, Andreas, Moneta Alessio (2011), "Back to Engel? Some evidence for the hierarchy of needs," Papers on Economics and Evolution 2011-13, Max Planck Institute of Economics, Evolutionary Economics Group.
Chai, Andreas, Moneta, Alessio (2012), Back to Engel? Some evidence for the hierarchy of needs. Journal of Evolutionary Economics (2012)
Engel E. (1857), Die Productions –und consumtions –Verhältnisse des Königreichs Sachsen , Bulletin de l’Institut International de Statistique9
Foellmi, R., Zweimuller, J., (2006), Income distribution and demand induced innovations, Review of Economic Studies 82, 95 112‐
Frenken K, van Oort FG,Verburg T (2007) Related variety, unrelated variety and regional economic growth, Regional Studies 41 (5): 685-697
Funke M, Ruhwedel R (2001a) Product variety and economic growth: Empirical evidence for the OECD countries, IMF Staff papers, 48, No. 2
Funke M, Ruhwedel R (2005) Export variety and economic growth in East European transition economies, The Economics of Transition 13(1): 25-50
Funke M, Ruhwedel R, (2001b) Export variety and export performance: empirical evidence from East Asia, Journal of Asian Economics 12: 493-505
Hidalgo CA, Klinger B, Barabasi A-L, Hausmann R (2007) The product space conditions the development of nations, Science 317(5837): 482-487
Imbs J., Wacziarg R., (2003), Stages of Diversification, American Economic Review, vol. 93, no. 1, p. 63-86.
Jevons WS (1924) The theory of political economy, 4th edn. MacMillan, London
Keynes J.M. (1928), Economic possibilities for our grandchildren, in Pecchi L., Piga G. (Eds), Revisiting Keynes: Economic possibilities for our grandchildren, Cambridge MIT Press (2008)
Lancaster K.J. (1966), A new approach to consumer theory, Journal of Political Economy, Vol. 14 (1966) 133-156.
Manig C., Moneta A. (2014), More or better? Measuring quality versus quantity in food consumption, J. Bioeconomics (2014) 16:155–178; DOI 10.1007/s10818-013-9174-4
Marshall A (1949), Principles of economics, 8th edn. MacMillan, New York
Matsuyama, K., (2002), The Rise of Mass Consumption Societies, Journal of Political Economy 110, 1035 70.‐
Menger C. (1950), Principles of economics, New York University Press, New York London
27
Nelson, R. Winter, S. (1982), An Evolutionary Theory of Economic Change, Cambridge, Mass, Harvard University Press, (1982).
Pasinetti L. L. (1981), Structural Change and Economic Growth. Cambridge, Cambridge University Press.
Pasinetti L.L., (1993), Structural Economic Dynamics, Cambridge, Cambridge University Press.
Pecchi L. Piga G. (Eds) (2008), Revisiting Keynes: Economic Possibilities for our Grandchildren, Cambridge MA, MIT Press
preferences. Industrial and Corporate Change 8: 353–399
Pyka, A., Saviotti, P.P. (2011), Economic Growth Through The Emergence of New Sectors, in Mann S. (Ed) Sectors Matter, Berlin- Heidelberg, Springer
Romer, P. (1990), Endogenous technical progress, Journal of Political Economy, Vol. 98, 71-102.
Saviotti P.P. (2001), Variety, growth and demand, Journal of Evolutionary economics, Vol.11 (2001) 119-142
Saviotti P.P., Metcalfe J.S. (1984) A Theoretical Approach to the Construction of Technological Output Indicators, Research Policy, 13 (1984), 141-151
Saviotti P.P., Pyka A. (2004), Economic development by the creation of new sectors, Journal of Evolutionary Economics, 14, (1) (2004) 1-35.
Saviotti P.P., Pyka A. (2008), Micro and macro dynamics: Industry life cycles,inter-sector coordination and aggregate growth, Journal of Evolutionary Economics, Vol 18 (2008) pp. 167-182
Saviotti P.P., Pyka A. (2012), On The Co-Evolution Of Innovation And Demand: Some Policy Implications, Revue de l'OFCE, n°124, pp 349-388 - Debates and policies: Agent-based models and economic policy, edited by Jean-Luc Gaffard and Mauro Napoletano
Saviotti P.P., Pyka A. (2013), From necessities to imaginary worlds: Structural change, product quality and economic development, Technological Forecasting & Social Change, Vol. 80 (2013) 1499–1512
Saviotti P.P., Variety, growth and demand, Journal of Evolutionary Economics (2001) 11: 119–142
Schumpeter, J. (1934, original edition 1912) The Theory of Economic Development, Cambridge, Mass, Harvard University Press.
Solow, R.M. (1956), A contribution to the theory of economic growth, Quarterly Journal of Economics, Solow, R.M., Vol 70 (1956) 65-94.
Solow, R.M. (1957), Technical change and the aggregate production function, Review of Economics and Statistics, Vol. 39 (1957) 312-320.
Stiglitz J.E. (2010), Toward a general theory of consumerism: Reflections on Keynes 's Economic possibilities for our grandchildren, in Pecchi L., Piga G. (Eds), Revisiting Keynes: Economic possibilities for our grandchildren, Cambridge MIT Press (2010)
Trigg, A. (2001), "Veblen, Bourdieu, and conspicuous consumption", Journal of Economic Issues, Vol. 35 No.1, pp. 99–115
Veblen, Thorstein B. (1899) The Theory of the Leisure Class: An Economic Study in the Evolution of Institutions (New York: Macmillan). Maslow, A. H. (1943). "A theory of human motivation". Psychological Review 50 (4): 370–396.
Walras L (1896) Elements d’economie politique Pure, 3rd edn. Economica, Paris (reprinted 1988)
Whaples, Robert (2001), “Hours of Work in U.S. History”. EH.Net Encyclopedia, edited by Robert Whaples. August 14, 2001. URL http://eh.net/encyclopedia/hours-of-work-in-u-s-history/
28
Witt U. (2001), "Learning to consume: a theory of wants and the growth of demand." Journal of Evolutionary Economics, Vol. 11.
Zilibotti F. (2010), Economic possibilities for our grandchildren 75 years after: a global perspective, in Pecchi L., Piga G. (Eds), Revisiting Keynes: Economic possibilities for our grandchildren, Cambridge MIT Press (2010).
29