triggers of change: structural trajectories and production dynamics

Cambridge Journal of Economics 2013, 1 of 18doi:10.1093/cje/bet034

© The Author 2013. Published by Oxford University Press on behalf of the Cambridge Political Economy Society. All rights reserved.

Triggers of change: structural trajectories and production dynamics

Antonio Andreoni and Roberto Scazzieri*

The transformation of production structures has been at the centre of the histori-cal dynamics of capitalist economies since the first Industrial Revolution. The arti-cle concentrates on production processes as principal loci of structural economic dynamics along increasing and decreasing returns trajectories. These trajectories are triggered by structural opportunities and constraints embedded in production systems, and their historical realisation is subject to different institutional configura-tions—production units at different levels of aggregation. The approach envisioned suggests the possibility of governing economic dynamics by structural policies working on the technological and organisational conditions of production as well as on the configuration of production processes within production units. Capitalist economies have to rely on a mix of co-ordination devices across different produc-tion units and aggregation levels for capturing structural opportunities and avoiding structural constraints.

Key words: Structural economic dynamics; Production structure; Increasing returns; Decreasing returns; Structural policiesJEL classifications: D24, L16, O25

1. Introduction

The transformation of production structures has been at the centre of the historical dynamics of capitalist economies since the first Industrial Revolution. Increasing and decreasing returns trajectories came to the fore, especially when accelerated growth during the first Industrial Revolution triggered decreasing returns due to demand expansion and non-produced resource bottlenecks, and increasing returns due to new production possibilities in manufacturing and quantitative/qualitative changes in demand composition. The aim of this article is to concentrate on the production domain as the principal locus of trajectories of structural change due to the unfolding of opportunities and constraints over time.

Manuscript received 15 February 2012; final version received 10 March 2013.Address for correspondence: Antonio Andreoni, University of Cambridge, Department of Engineering,

Institute for Manufacturing, 17 Charles Babbage Road, Cambridge, CB3 0FS, UK; email: [email protected]

* Centre for Science, Technology and Innovation Policy, Institute for Manufacturing, Department of Engineering, University of Cambridge (AA) and Department of Economics, University of Bologna and Gonville and Caius College and Clare Hall, Cambridge (RS).

Cambridge Journal of Economics Advance Access published July 19, 2013 at U

niversity of Cam

bridge on August 11, 2013

http://cje.oxfordjournals.org/D

ownloaded from

mailto:[email protected]

http://cje.oxfordjournals.org/

of 18 A. Andreoni and R. Scazzieri

Smith’s increasing returns and Malthus-Ricardo’s decreasing returns, respectively, highlight i) opportunities that can materialise within the production system through division of labour, achieved through specialisation and activation of complementarities (increasing returns) and ii) constraints that have to be faced through the change of pro-duction structures if the economic system is to continue along a self-sustained path of accumulation and expansion (decreasing returns). In our view, trajectories of feasible transformations made possible by opportunities and constraints in production struc-tures should be distinguished from the trajectories followed by unfolding increasing and decreasing returns under given historical conditions. This approach allows identifica-tion of the fundamental causal structure behind increasing and decreasing returns and suggests its generalisation to historical contexts different from those in which they were originally detected. In other words, we assume that historical processes are inherently multi-layered, so that disentangling relatively persistent from more contingent features of historical reality may be useful in extracting causal principles from specific contexts and applying them when examining the current and future prospects of capitalist economies.

The distinction between structural opportunities/constraints and actualised possibili-ties in historical contexts suggests an important set of questions: i) what can we learn about the specific technological and organisational conditions triggering the unfolding of opportunities or the prevention of constraints? ii) How might those conditions may differ according to the level of aggregation of production processes? iii) How can we identify specific policy measures capable of driving economies along historical trajecto-ries in which increasing returns are triggered and/or decreasing returns are prevented?

Policy measures effective in triggering increasing returns and/or preventing decreas-ing returns are bound to be relevant at two different operational levels: i) the achieve-ment of certain technological and organisational conditions making increasing and/or decreasing returns possible and ii) the achievement of certain institutional conditions allowing realisation of increasing returns and/or prevention of decreasing returns within specific production units. This distinction is central to the selection of which policy mix may be considered as the most effective in any given context. For instance, when increasing returns are possible due to existing technological complementarities between production processes, policy measures acting on co-ordination across different produc-tion units may be most effective. On the other hand, contexts in which technological complementarities are not present may suggest policy measures acting on the internal structures of production processes so as to trigger complementarities across processes.

This article explores the multi-layered structure of production arrangements and maintains that the distinction between different orders of change may be essential to identify causal explanations and effective policy strategies. Section 2 frames our contribution in the literature of structural economic dynamics and relates structural change as a change in sectoral proportions with change within production processes. Sections 3 and 4 investigate increasing and decreasing returns trajectories character-ising the historical dynamics of capitalist economies. This is done by highlighting the multiple triggers of change at play, that is by identifying structural opportunities and constraints embedded in production systems and by pointing out that their realisa-tion is subject to different institutional configurations—that is, production units at different levels of aggregation—in any given historical context. The implications of our approach for the design of structural policies and the selection of an effective mix of co-ordination devices are discussed in Section 5. Concluding remarks bring the article to close.

at University of C

ambridge on A

ugust 11, 2013http://cje.oxfordjournals.org/

Dow

nloaded from


Triggers of change of 18

2. Production dynamics and the dynamics of capitalism

Classical political economy examines structural economics dynamics in terms of laws conducive to a variety of outcomes depending on historical triggers (Lowe, 1954, pp. 157–58). Smith views an economic system subject to growing extent of the market and increasing division of labour, whereas Malthus-Ricardo emphasise the long-term influence of non-produced essential inputs (such as land and mines). In both cases, structural change trajectories work themselves out through qualitative and composi-tional changes and according to relatively invariant causal structures. Both Smith and Ricardo identify trajectories in which the transformation of production arrangements leads to changes in relative proportions between productive sectors. In Smith (1976 [1776]), division of labour increases at different rates in different productive activities (say, it increases at a higher rate in manufacturing relative to agriculture) and is likely to lead to different maximum growth rates across activities. In Ricardo (1951 [1817]), the limited availability of land and mines leads to the use of less effective technical practices as activity levels are increased. In this case the proportions among activities are changed as the economic system moves from more to less effective lands (or mines) and/or from less to more intensive cultivation (or extraction) through the introduction of more capital intensive crop-growing (or extraction) techniques.

Modern theories of structural economic dynamics often emphasise the relationship between the changing composition of production and economic expansion of the over-all economic system (Robinson, 1956; Kaldor, 1967, 1981; Leon, 1967; Lowe, 1976; Pasinetti, 1981, 1993; Quadrio Curzio and Pellizzari, 1999).1 The constitutive role of compositional changes of aggregate production is also emphasised in the empirical and historical literature on modern economic growth and development trajectories (Hirschman, 1958; Chenery, 1960; Furtado, 1964; Chenery and Taylor, 1968; Cole and Deane, 1969; Kuznets, 1971; Syrquin, 1988). Here, the underlying set of possible causal relationships is not systematically understood as resulting from the internal transformation of production structures.

In Joan Robinson’s The Accumulation of Capital, technical progress and demographic growth are at the root of expansionary paths associated with changing proportions amongst production sectors (Robinson, 1956). However, the internal dynamics of the production system is not seen as the fundamental reason behind the specific configura-tion taken by the economic system at any given time (see also Harcourt and Kerr, 2009, pp. 76–100, 118–40). In a similar vein, Luigi Pasinetti’s Structural Change and Economic Growth takes technical progress and changing consumers’ habits (via Engel’s law) to be the determinants of changing proportions among production sectors as required by mac-roeconomic equilibrium conditions (Pasinetti, 1981). Here, the analytical focus is on satisfying the latter conditions through specific structural dynamics of the production sys-tem rather than investigating the specific causal systems behind those dynamics (see also Harcourt, 2006, pp. 123–33). Nicholas Kaldor’s Strategic Factors of Economic Development and Adolph Lowe’s Path of Economic Growth take a different view insofar as long-term economic dynamics are seen as resulting from growth processes taking place within a par-ticular productive sector such as manufacturing (Kaldor, 1967) or from responses within

1 In particular, changes of sectoral proportions associated with macroeconomic dynamics may derive from the different growth rates of sectors in the economic system (Landesmann and Scazzieri, 1990; see also Baranzini and Scazzieri, 1990 and Scazzieri, 2009).

at University of C

ambridge on A


Dow

nloaded from



the production system to impulses external to that system (Lowe, 1976). In a differ-ent but complementary perspective, Alberto Quadrio Curzio has investigated decreasing returns trajectories triggered by the limited availability of non-produced resources and has highlighted that complementarities among production processes may result in the possibility of temporary increases of productive efficiency on a decreasing returns trajec-tory (Quadrio Curzio, 1975, 1986; Quadrio Curzio and Pellizzari 1999).

The following section outlines a reformulation of theories of increasing and decreas-ing returns in terms of the above-mentioned distinction between opportunities/con-straints in production structures and actualised production possibilities in historical contexts. This point of view has close resemblance to the ‘separation theorem’ recently introduced by Luigi Pasinetti insofar as we also distinguish between structural and institutional features of production activity (Pasinetti, 2007, p. 275). In our view, how-ever, the structural sphere consists of two domains (the technological and the organi-sational domains, respectively) joined together by means of institutional arrangements (production units at different levels of aggregation). Institutions are thus the necessary route by which structural opportunities and constraints may or may not be realised.

3. Disentangling increasing and decreasing returns: structural trajectories for the long run

Classical and modern theories of structural dynamics have addressed increasing and diminishing returns without fully investigating the causal principles underlying the transformation of production structures. If we look at increasing returns through the lenses of Smith’s classical trio of advantages, we discover that increasing returns are generated through a plurality of organisational and cognitive causes whose interaction is not directly investigated by Smith (see Scazzieri, 2013):

This great increase of the quantity of work which, in consequence of the division of labour, the same number of people are capable of performing, is owing to three different circumstances; first to the increase of dexterity in every particular workman; secondly, to the saving of the time which is commonly lost in passing from one species of work to another; and lastly, to the inven-tion of a great number of machines which facilitate and abridge labour, and enable one man to do the work of many. (Smith, 1976 [1776], I.1.5)

A more detailed causal reconstruction of increasing returns is to be found in Charles Babbage’s treatment in terms of a ‘law of multiples’ within production structures: ‘When the number of processes into which it is most advantageous to divide produc-tion, and the number of individuals to be employed in it, are ascertained, then all factories which do not employ a direct multiple of this latter number, will produce the article at a greater cost’ (Babbage, 1835, p. 213). Babbage’s treatment suggests a struc-tural analysis of increasing returns that contemplates indivisible processes and a law of multiples, the latter implying a criterion of proportionality. In fact, attaining a certain minimum process scale is a necessary condition for the introduction of more effective production techniques. At the same time, the law of multiples calls attention to the fact that maintaining any given level of productive efficiency presupposes increasing pro-cess scale by integer multiples of the minimum scale allowing that level of efficiency. As a result, a continuously increasing process scale may allow any given productive establishment to shift from lower to higher levels of efficiency if the minimum scale requirements corresponding to any given level of efficiency are satisfied.

at University of C

ambridge on A


Dow

nloaded from



The classical approach to decreasing returns points to multiple interacting causes. A clear instance of this is the treatment of decreasing returns in Malthus’s pamphlet on the Nature and Progress of Rent:

Every extensive country may thus be considered as possessing a gradation of machines for the production of corn and raw materials, including in this gradation not only all the various qualities of poor land, of which every large territory has generally abundance, but the inferior machinery which may be said to be employed when good land is further and further forced for additional produce. As the price of raw produce continues to rise, these inferior machines are successively called into action; and as the price of raw produce continues to fall, they are succes-sively thrown out of action. (Malthus, 1815, p. 39)

Here decreasing returns derive from resource availability and a specific mode of resource appropriation. However, only recently has the duality between structural conditions and institutional arrangements along decreasing returns trajectories become a focus of atten-tion. In particular, the classical concept of natural order of land fertility has been challenged by acknowledging the influence of prices and income distribution on the efficiency ranking of lands (Sraffa, 1960, chapter XI; Quadro Curzio, 1967; Quadrio Curzio and Pellizzari, 1999). The acknowledegment of this fallacy in the classical theory of decreasing efficiency is made possible by the distinction between structural constraints and institutional con-figurations of production structures. A non-producible input such as land introduces spe-cific upper bounds for any given production technique. However, institutional conditions determine which particular techniques will be adopted for any given scale of production. This distinction between structural and institutional conditions is a powerful heuristic for disentangling the causal structures underlying the dynamics of capitalist economies.

In the case of increasing returns, neither Smith nor Babbage presuppose production units at a specific level of aggregation. They both discuss the advantages of division of labour by referring to the allocation of equipment’s production capacities and human capabilities to tasks within large productive establishments (Smith’s pin manufacture or Babbage’s factory), but their respective arguments do not presuppose a one-way relationship between increasing returns and the economies of large production. Indeed, Smith’s relationship between an increasing scale of production and the increasing divi-sion of labour is compatible with multiple levels of aggregation of productive activities and multiple sizes of productive units for each level of aggregation (Young, 1928; Ames and Rosenberg, 1967). A similar argument applies to Babbage who, after stating his ‘law of multiples’ (see previous discussion), emphasised that the pattern of division of work ‘which is most profitable for a factory employing a hundred workmen, may not be quite the best where there are five hundred; and the arrangements of both may probably admit of variations, without materially increasing the cost of their produce’ (Babbage, 1835, p. 212, emphasis added). In short, classical theories of increasing returns highlight the structural conditions for improvement of productive effectiveness without committing themselves to any specific choice of aggregation level and size of production unit. This approach is consistent with the view later expressed by Alfred Marshall when he wrote that ‘the economies of production on a large scale can seldom be allocated exactly to any one industry: they are in great measure attached to groups, often large groups, of cor-related industries’ (Marshall, 1920 [1919], p. 188, emphasis added).2

2 Marshall’s attention for the advantages of correlated industries is also at the root of his theory of external economies. In fact, economies of large-scale production and external economies are two distinct effects of the same set of causal interdependencies.

at University of C

ambridge on A


Dow

nloaded from



Decreasing returns are similarly associated with the distinction between structural conditions and institutional contexts, as pointed out in contributions subsequent to Malthus-Ricardo’s formulation of decreasing returns (Antonelli, 1910; Byé, 1928). In particular, Piero Sraffa called attention to the relationship between the specification of constraints and the level of aggregation of productive activities:

If every single industry is defined as the exclusive consumer of a given factor of production (for example, agriculture, the iron industry, etc.), a condition is at once assumed that implies a ten-dency for increasing costs for the industry, since it is precisely the factor that is characteristic of the industry (cultivable land, iron mines, etc.) that, with the increase of production, generally remains constant. If, on the contrary, every industry is defined as the sole producer of a given product, and this is meant in a very restrictive sense, so that in general it can be thought that every industry uses only a small fraction of each factor of production (negligible in comparison to the quantity used by all the industries together), we thereby exclude from the industry the circumstance that generate increasing costs. (Sraffa, 1998 [1925], p. 357)

The reason the central causal structure generating decreasing returns depends on the definition of the production unit is because different resource bottlenecks mate-rialise at different levels of aggregation. To sum up, trajectories of structural change associated with increasing and decreasing returns are triggered by opportunities and constraints both internal and external to production processes. Given those triggers, both increasing and decreasing returns operate through ‘interlocking and cumulative’ processes of reconfiguration of production structures (Myrdal, 1958, p. 23).

4. Structural opportunities and constraints in production units

The investigation of what makes any given production system conducive to specific trajectories of structural change rather than others presupposes a detailed assessment of the different structural components of any given production process and their rela-tionships. Production processes may be represented as systems of inter-related tasks through which a sequence of transformations of materials are performed according to different combination of flow and fund inputs, subject to certain scale and time constraints (Georgescu-Roegen, 1970; Morroni, 1992; Scazzieri, 1993; Landesmann and Scazzieri 1996; Andreoni, 2013). The distinction between fund inputs (such as productive agents and mechanical artefacts) and flow inputs (such as fuel, chemical catalysts and electricity) calls attention to the degree of permanence of these inputs in the entire production process. Fund inputs maintain their characteristics substan-tially unaltered during a production process, provided certain tolerance thresholds are not violated (Landesmann, 1986). Mechanical artefacts present certain production capacities, whilst each productive agent is characterised by a certain set of production capabilities.

Different production processes are associated with different configurations of tasks, capabilities/capacities and materials; also, the same production process may be per-formed according to different configurations of its components. Opportunities for structural change, or constraints on it, are specific to the ways any given process results from those specific configurations. In particular, alternative precedence patterns among causal factors may trigger different trajectories of increasing or decreasing returns, within and across industries. For example, opportunities arising in certain mechanical industries conducive to widespread division of labour may be a necessary condition for the realisation of opportunities within other industries such as mechanically induced

at University of C

ambridge on A


Dow

nloaded from



improvements in crop-growing processes—that is, inter-sectoral learning triggered by complementarities (Rosenstein-Rodan, 1943; Andreoni, 2011). Two fundamental ‘production facts’ are at the root of the existence of a plurality of ‘worlds of production’ (Salais and Storper, 1997).

First, any given production process involves the intertwining of technological and organisational domains. In the technological domain, the set of relevant tasks to be performed as well as their characteristics are defined according to i) specific physical, chemical and biological laws regulating the production process in terms of time, scale and proportionality requirements; and ii) the specific properties of materials in use. These laws and properties imply specific conditions of task feasibility. In other words, they define the limited set of structurally feasible tasks and combinations of tasks that have to be performed for obtaining a certain productive output.3 In the organisational domain the set of structurally feasible tasks are assigned to productive agents (each of them complemented by mechanical artefacts) according to specific co-ordination criteria and behavioural principles. A plurality of organisational configurations is pos-sible, given the set of structurally feasible tasks determined within the technologi-cal domain. Thus, any given process of production may be described in terms of the duality between the domain of structurally feasible tasks and that of organisational arrangements.

The second ‘production fact’ is that the transformation of materials takes place within production units that may be identified at different levels of aggregation, such as the productive establishment, the constellation of establishments, the industry or the production system as a whole (Marshall, 1920 [1919]; Becattini, 1989). Production units are not only classificatory devices but point to the existence of agents and/or col-lections of agents performing production processes at a specific level of aggregation. Each production unit is endowed with specific collections of capacities/capabilities whose activation and transformation are the outcome of the interplay between cogni-tive dynamics and production structures (Richardson, 1972; Pitelis and Teece, 2009; Jacobides and Winter, 2012).4

Any given production unit rests on the interaction between the technological and organisational domains. It follows that the transformation of production structures may result from the interaction between different sources of change. On the one hand, technological or organisational changes are possible within any given unit of produc-tion. On the other hand, changes in production units may trigger transformations in the technological and organisational structures of production. Figure 1 shows the rela-tionship between different representations of any given production process, where T is the domain of structurally feasible tasks, O is the set of organisational arrangements also known to be feasible and uk is the generic production unit.

3 The interplay between cognitive schemes and production structures is discussed in Andreoni (2013). Here the concept of ‘structural learning’ describes a process of learning (and thus capacities/capabilities transformation) that is embedded and oriented by complementarities, similarities and indivisibilities within production structures. For instance, the transformation of capacities/capabilities along an increasing returns trajectory pre-supposes attainment of a certain minimum scale subject to a given law of proportionality amongst components of the production system (Scazzieri, 2013).

4 As Rosenberg (1963, p. 440) notes, ‘an analytical explanation of many of the technological changes in the manufacturing sector of the economy may be fruitfully approached at the purely technological level. This is not to deny, of course, that the ultimate incentives are economic in nature; rather, the point is that complex technologies create internal compulsions and pressures which, in turn, initiate exploratory activity in particular directions’.

at University of C

ambridge on A


Dow

nloaded from



The interfaces between the two domains and along the different levels of aggre-gation allow identification of a hierarchy of structural opportunities and constraints. Trajectories of structural change take different forms depending on whether we con-sider i) the technological or organisational opportunities and constraints associated with production units at a certain level of aggregation, or ii) the opportunities and constraints associated with productions unit as we move from one level of aggregation to another. For instance, as we expand the size of any given constellation of interact-ing establishments, it is likely that new opportunities for specialisation of productive activities and increasing returns would arise (Young, 1928). But it is also likely that as network size is increased, additional scarcity constraints will be discovered because an increasing number of non-produced inputs will be essential and exclusive providers of productive services to the expanded network are likely to appear. On the other hand, a resource such as iron ore that appears to be free at a given level of aggregation may turn out to be limited when the level of aggregation is changed. This condition may be expressed by considering the following matrix of physical upper bounds on the absorp-tions of non-produced resources (see Scazzieri, 1993, pp. 196–97).

Z

Z Z

Z

Z

Zij

nk

=

11 12

22

0 0

0 0 0

0 0 0

0 0 0

Each element of matrix Z denotes the upper bound on absorptions of non-produced resource i into the process (or industry) delivering product j. In our example, three non-produced inputs are process-specific and the corresponding upper bounds are, respectively, z22, zij and znk. The argument entails that production units are central to

Fig. 1. A multi-layered representation of production activity.

at University of C

ambridge on A


Dow

nloaded from



the generation of different trajectories of structural change. Increasing and decreas-ing returns may be generated within different production units at different levels of aggregation. Indeed, the switch from one level of aggregation to another should be dis-tinguished from the switch from one size of productive unit to another. For instance, switching from the individual establishment to a network of establishments does not necessarily lead to a switch from small to larger establishments, insofar as any given network of establishments may be compatible with a plurality of establishment sizes.

Decreasing and increasing returns are triggered by changes in the scale of production affecting the relationship between organisation and technology within any given unit of production. However, the relationship itself may take a different form depending on which unit of production (and level of aggregation) we are considering. In the decreas-ing returns case, for any given size of production unit uk an increase in the scale of pro-duction, as measured by the corresponding output level(s), beyond a certain threshold brings about decreasing returns. However, an increase in the size of the production unit beyond uk may take the production structure back to its pristine state, insofar as the pro-duction unit under consideration is able to increase its endowment of the non-produced resource(s) responsible for the onset of decreasing returns. In the increasing returns case, lower bounds on the minimum scale of production compatible with upwards jumps in productive efficiency are central to the unfolding of an increasing returns trajectory. For any given size of the production unit uk, any increase in the scale of production beyond the minimum scale compatible with maximum production efficiency according to the Babbage law of multiples brings down production efficiency, unless the size of the production unit is increased by integer multiples of the minimum size compatible with maximum efficiency. Furthermore, scale and size increases make it feasible to bring back the previous production structure and level of efficiency. However, it is also possible that a sufficiently high increase in the scale of production would make it feasible to introduce a new (and more effective) combination of technology and organisation, thereby allow-ing the further unfolding of the increasing returns trajectory.

The relationship between the scale and the size of production units along decreas-ing and increasing returns trajectories may be illustrated within the multi-layered representation of production activity we have introduced (see Figure 1). Given cer-tain organisational and technological conditions, the matrix that follows represents a decreasing returns process within a sequence of different units of production.

o o o o o o o o o ot

t

t

t

t

t

t

t

t

t

u

u

u

u

u

u

u

1 2 3

1

2

3

1 2 3 1 1 2 1

1

2

3

1

1

2

1

1

1

1

2

2

2

Scale

22

3

3

4

u

u

u

size

at University of C

ambridge on A


Dow

nloaded from



The matrix shows that an increasing size of the productive unit (from u1 to u2, from u2 to u3, and from u3 to u4) is associated with discontinuous changes of both produc-tion technology and production organisation, and this process may take place at the same time we observe a continuous increase of the scale of production. In particular, the matrix shows that a decreasing returns trajectory may be stopped and reversed as the productive size is increased from u1 to u2. At u2, an increasing scale of production again triggers decreasing returns until the productive size is increased from u2 to u3 and so on.

Given certain organisational and technological conditions, the next matrix repre-sents an increasing returns process within a sequence of different units of production.

o o o o o o o o o ot

t

t

t

t

t

t

t

t

t

u

u

u

u

u

u

1 2 3 1 2 3 4 1 2 1

1

2

3

1

2

3

4

1

2

1

1

1

1

2

2

*

*

Scale22

2

3

3

4

u

u

u

u

size

The matrix shows that an increasing scale of production associated with a given size of the productive unit u1 is bound to lead to changes of production technology and pro-duction organisation due to failure to satisfy the law of multiples. As the size of the pro-ductive unit rises to u2, an increasing scale of production may again be compatible with the law of multiples, and the production structure may revert to the original level of effi-ciency. Further increases of the scale of production are associated with lower efficiency unless the law of multiples is satisfied (in which case the technology- organisation pair allowing maximum efficiency can be maintained through time). The shift to productive size u4 allows an upward jump of productive efficiency associated with technology t1* and organisation o1*. The matrix shows that the scale of production may be continu-ously increasing whilst productive efficiency is subject to upward and downward jumps.

5. Structural policies: which way forward?

Opportunities and constraints embedded in production structures are not always actu-alised or, respectively, binding: the space of production possibilities that are structur-ally feasible has to find its ‘historical closure’ in specific production units operating in different contexts. As we have seen, opportunities and constraints may be highly ‘interlocking’, that is, overcoming certain constraints unlocks opportunities embed-ded in the production structure, whereas capturing certain opportunities may set the production structure along a path characterised by new constraints. This type of

at University of C

ambridge on A


Dow

nloaded from



interdependence changes according to the different production units (e.g., productive establishment, the constellation of establishments, the industry or the production sys-tem as a whole) at the different levels of aggregation of production activities. This point of view makes it possible to conceive selective interventions that would target specific sets of production units at different level of aggregation so as to allow the realisation of opportunities that would otherwise not be taken (this is the case of a ‘missed’ increas-ing returns trajectory) and/or the identification of constraints that would otherwise go unnoticed (this is the case of a ‘prevented’ decreasing returns trajectory). Interventions of such kind are labelled structural policies.

The conceptual framework already outlined highlights that certain trajectories of structural change presuppose satisfaction of structural conditions determined by the technological and organisational characteristics specific to the production activities selected. Moreover, as these structural conditions tend to vary according to the pro-duction unit considered, policies will differentiate according to the unit of analysis and the structural conditions under consideration.5 This implies that there is no structural policy as such but multiple structural policies can be designed and implemented. The possibility of policies acting on specific sets of production activities to achieve certain outcomes desirable for the socio-economic system as whole has also been acknowl-edged both by classical (List, 1885 [1841]) and modern analyses of industrial policies (Chang, 1994, 2002; Ocampo et al., 2009). For example, industrial policies may typi-cally allow the economic system to achieve certain structural conditions of proportion-ality at the level of the industry, which the single productive establishments would not be able to achieve autonomously or in the form of a constellation of establishments.

The perspective outlined in this article raises the possibility of designing selective structural policies because this analytical framework allows acting on different tech-nological and organisational domains and different production units through different policy measures. As a result structural policies may be implemented by different actors such as national, regional and local governments as well as entrepreneurial organisa-tions. The ‘production unit’ perspective envisioned here provides a focussing device in the identification of those actors who are most capable to give direction to the complex processes in which i) opportunities may be discovered and captured, and ii) constraints may be identified and avoided. Historical examples have shown how public actors are differently capable of establishing a web of embedded autonomy relation-ships with productive actors operating in productive units at different levels of aggrega-tion (Amsden 1989; Wade, 1990; Chang, 1994, 2002; Evans, 1996).

Given a certain socio-economic system and different policy actors, the visualisation of the structural conditions specific to any existing production process in different production units at different levels of aggregation requires specific industrial diag-nostics (Andreoni, 2012). The latter can be applied either as a heuristic of structural feasibility or as a heuristic of contextual viability in the design of structural policies.

5 One important related issue is that different time horizons may be associated with different thresholds for increasing returns, or different bottlenecks for decreasing returns. A longer time horizon may be associ-ated with more degrees of freedom in terms of overcoming thresholds along an increasing returns trajectory and/or in terms of avoiding non-produced resource bottlenecks along a decreasing returns trajectory. This time dependence of increasing and decreasing returns trajectories has critical implications for what concerns the selection of policy measures. Policy choices are especially constrained in the short run (when policies must deal with the ‘step coming next’) but are less and less so as one moves to longer time horizons (Sylos Labini, 1970, pp. 15–16).

at University of C

ambridge on A


Dow

nloaded from



This distinction arises from the observation of countries at different levels of develop-ment, and thus characterised by different production structures. For each country, given a certain historical context, one can identify productive configurations that are structurally feasible but not contextually viable, and productive configurations that are contextually viable but not yet actualised.

As a heuristic of structural feasibility, industrial diagnostics are intended to visu-alise the technological and organisational arrangements that are feasible in differ-ent production units at different levels of aggregation. In other words, they provide a structural benchmark for policy design. This benchmark allows discovery of which specific opportunities and constraints may arise under given institutional set-up—that is, productive units—as we move from one level of aggregation to another. However, visualising the set of opportunities as well as constraints that the satisfaction of certain structural conditions makes feasible does not directly inform policy actors about what is possible in the specific context they operate—that is, contextual viability condition.

As a heuristic of contextual viability, industrial diagnostics are intended to provide a structural picture of the specific historical conditions characterising a certain socio-eco-nomic system at a certain moment in time. In any historical context, a socio-economic system is characterised by a certain endowment of mechanical capacities and human capabilities as well as it is conditioned by external constraints due to resource availabil-ity. Human capabilities and mechanical capacities are embedded in production units at different levels of aggregation and accumulate through processes of transformation which realise at the level of both the technological and organisational domains. Given a certain production output, human capabilities, mechanical capacities and materials in use (as well as their organisation) define the set of technical practices.

Given a socio-economic system at a certain stage of development, Figure 2 visualises the different spaces just considered, namely, the space R of historically realised pro-duction processes, the space V of production processes that are contextually viable but not historically realised and the space F of production processes that are structurally feasible but not viable given the historical context V. The generic trajectories vti and ftj represent, respectively, the contextually viable path i and the structurally feasible path j that a given socio-economic system (and production processes performed in time pt as its components) can experience starting from a given historical context. For each economic system, the possibility of going through any contextually viable path i and that of entering a structurally feasible path j are strongly inter-dependent. Indeed,

Fig. 2. Structurally feasible and contextually viable trajectories.

at University of C

ambridge on A


Dow

nloaded from



complementarities among the trajectories vti and ftj imply an order of precedence in structural change processes and tend to result in unequal dynamics in quantities, prices and sectoral employment.

Every time an economic system enters a certain path, both the space of contextually viable production possibilities V and the space of structurally feasible configurations F mutate, both qualitatively and quantitatively. As a matter of fact, one may conjec-ture that in the long run the opportunities within space V will gradually materialise, whilst opportunities that were only structurally feasible become contextually viable. Certain socio-economic systems may actually directly ‘jump’ from space R to space F without fully exploiting all the opportunities available within V (this would be a typi-cal case of ‘leapfrogging’). As discussed, the trajectories vti and ftj can be triggered by realising increasing returns through the activation of suitable production units and/or co-ordination devices, or by avoiding constraints that may generate decreasing returns.

In the increasing returns case, given a certain endowment of human capabilities and mechanical capacities, it is generally possible to recombine production tasks and organisational arrangements within the T and O domains so as to satisfy the relevant proportionality conditions by moving from one level of aggregation to another. This allows the actors to fully explore the range of alternative technical practices that are structurally feasible and to choose among those alternatives those that are contextually viable under given historical conditions.

In the decreasing returns case, given external constraints due to resource availability, it is often possible to move from one level of aggregation to another to avoid the actual onset of decreasing returns. However, this possibility depends on whether we are deal-ing with ‘produced’ or ‘non-produced’ resources. For example, decreasing returns due to the limited availability of produced resources at the level of individual establishments may be avoided by moving up to the level of an establishment network (not necessarily an industry) and splitting up the original production process into a number of distinct and inter-related processes. Alternatively, decreasing returns due to the limited nature of non-produced resources at the level of an aggregate of industries, or of a single indus-try, may be avoided by moving down, respectively, to the level of individual industries or individual establishments (Sraffa, 1998 [1925], p. 357; see Section 2). In this case, the trigger of decreasing returns will be removed as long as it is possible to shift the limited resource from one industry to another (or from one establishment to another).

Triggers of structural dynamics can be very different from traditional instances of increasing or decreasing returns such as Schumpeterian innovations and technological transfer on the one hand, and technological decumulation of human capabilities—that is, de-skilling—or mechanical capacities in the case of manufacturing shrinking on the other hand.

Stylised historical facts support the idea that socio-economic systems have under-taken increasing (or decreasing) returns trajectories thanks to the working of mar-ket forces along with the implementation of structural policies by different actors (Rosenberg, 1969; Noble, 1984; Chandler, 1990, 2005; Simon, 1991; Sabel and Zeitlin, 1997; Chang, 2002; Chang et al., 2013). Even in capitalist systems, meant as a type of socio-economic systems in which private ownership of liquid funds is associ-ated with the use of those funds in productive activities (Hicks, 1969), different actors along with market forces have shaped the capitalist form of production allowing for the realisation of certain trajectories of structural change instead of others. Governments have not only supplemented market co-ordination failures, but also in certain historical

at University of C

ambridge on A


Dow

nloaded from



circumstances allowed socio-economic systems to meet those structural conditions and degree of concentration of production processes (market structure) necessary for increasing (decreasing) returns trajectories to work out (be avoided).

Combinations of market and non-market arrangements is, thus, typical of economic co-ordination under capitalist conditions (Coase, 1937; Richardson, 1972; Powell, 1990; Hall and Soskice, 2001; Ménard, 2004; Lazonick, 2007). Clearly, structural dynamics will be different depending on which co-ordination devices are in place amongst dif-ferent production units at different levels of aggregation. For this reason the alternative futures of capitalisms as systems of production are closely intertwined with the way different co-ordination devices are adopted differently at the level of establishment net-works or industries and across industries of any given economic system.

Different co-ordination devices are normally available, on the one hand, within any given production unit and, on the other hand, across production units at different levels of aggregation. For example, the individual establishment may introduce direct co-ordination and planning that are not normally available at the higher levels of aggre-gation (Penrose, 1959; Chandler, 1977); co-ordination within an industrial district is likely to require different devices from co-ordination within a non-localised industry or a large aggregate of industries (Becattini, 1989). Co-ordination within establishment networks, industries, or aggregates of industries normally relies upon a mix of market and non-market devices that is highly context-dependent: customary arrangements, union structures, informal co-ordination and formal governance complement and give direction to market structures that are often distant from the competitive model (O’Sullivan et al., 2013).

The most effective mix of co-ordination devices along an increasing returns trajec-tory would be one allowing division of labour and specialisation of industries accord-ing to the proportionality conditions specific to any given production unit and across production units for any given level of aggregation. This means that the distribution of human capabilities and mechanical capacities to production tasks, realising in the technological and organisational domains, will have to be compatible with the specific proportionality requirements holding for the particular production unit under con-sideration. It is reasonable to conjecture that different co-ordination devices would be required to achieve the latter objective for different production units. For example, one may conjecture that switching from the establishment to the industry would be compatible with a variety of co-ordination criteria from the transfer of inputs through exchange to mergers and acquisitions. On the other hand, switching from the indus-try to the establishment may require the administrative fixing of utilisation quotas (O’Sullivan et al., 2013).

The two extreme cases would be those of an ideal market economy where the full set of production activities is co-ordinated through the price mechanism independently of the different levels of aggregation and the one of a planned economic system, where co-ordination is achieved through production planning at all levels of aggregation (see, respectively, Arrow, 1974; Nurske, 1953; Chakravarty, 1987). A change in the mix of co-ordination devices along with changes in the above-mentioned structural condi-tions (technological and organisational) and institutional conditions (production units at different levels of aggregation) may allow any given socio-economic system to follow specific historical trajectories rather than others. In conclusion, the structural approach envisioned here suggests a governance of structural dynamics realised through the adoption of policies working both on the technological and organisational conditions

at University of C

ambridge on A


Dow

nloaded from



of production as well as on the configuration of production processes within produc-tion units. Different levels of aggregation of production activities may require different co-ordination devices to realise increasing returns possibilities and/or to prevent the onset of decreasing returns. As a result, this approach shows that in any given histori-cal context capitalist economies have to rely on a mix of co-ordination devices across different production units and levels of aggregation to capture structural opportunities and avoid structural constraints.

6. Concluding remarks

This article has addressed alternative trajectories of long-run structural dynamics by going back to the approach and conceptions of the classical economists and taking advantage of the work of those modern economists who have addressed long-term economic change from the point of view of a theory of structural economic dynam-ics. This has led us to emphasise the importance of production arrangements and the dynamics of production structures and to look at long-term economic changes by proposing a separation between structural and institutional levels of investigation. The two classical laws of increasing and decreasing returns epitomise the classical approach to the long-term economic dynamics of capitalism. The article has out-lined a reformulation of the two laws by adopting a dual representation of production structures and emphasising that different increasing and/or decreasing returns trajec-tories may be identified depending on which production units and levels of aggrega-tion are considered.

The plurality of structural trajectories that any given economic system may fol-low suggests that there is a critical role for institutional variety (a long-run property) and for targeted interventions (a short- or medium-run property). Our argument has shown that triggers of increasing and decreasing returns work themselves out through relatively persistent production structures in different ways depending on which level of aggregation and production units are considered. It has also shown that capitalist modes of production are intrinsically dependent on the working of different co-ordina-tion devices for any given level of aggregation and production unit.

To conclude, this article draws attention to the critical contribution that the analysis of the relationships between production structure and economic dynamics can provide in addressing two distinct but closely connected problems: i) to identify how triggers of change work their way through structurally constrained paths, and ii) to design tar-geted and selective measures aimed at setting the economic system on one or another of those transformation paths. It has been found that enabling increasing returns and/or avoiding decreasing returns may require shifting the whole economic system from one set of structural and institutional conditions to another and that to achieve the lat-ter objective structural policy at a plurality of levels may be necessary.

Bibliography

Ames, E. and Rosenberg, N. 1967. The progressive division and specialization of industries, in Purdue Faculty Papers in Economic History, 1956–1966, Homewood, IL, Richard D. Irwin, 345–62

Andreoni, A. 2011. ‘Manufacturing agrarian change: inter-sectoral learning, agricultural pro-duction and technological capabilities,’ DRUID Working paper series no. 11–13

at University of C

ambridge on A


Dow

nloaded from



Andreoni, A. 2012. ‘Productive capabilities indicators for industrial policies design,’ UNIDO Working paper series 17/2011

Andreoni, A. 2013. Structural learning: embedding discoveries and the dynamics of production, Structural Change and Economic Dynamics, forthcoming

Antonelli, E. 1910. Note sur la loi du rendement non proportionnel, Revue d’économie politique, vol. 24, July, 532–45

Amsden, A. 1989. Asia’s Next Giant: South Korea and Late Industrialisation, Oxford, Oxford University Press

Arrow, K. 1974. The Limits of Organization, New York, NortonBabbage, C. 1835. On the Economy of Machinery and Manufactures, 4th ed., London, Charles

KnightBaranzini, M. and Scazzieri, R. 1990. Economic structure: analytical perspectives, in Baranzini,

M. and Scazzieri, R. (eds.), The Economic Theory of Structure and Change, Cambridge, Cambridge University Press, 227–333

Becattini, G. 1989. Sectors and/or districts: some remarks on the conceptual foundations of industrial economics, in Goodman, E. and Bamford, J. with Saynor, P. (eds.), Small Firms and Industrial Districts in Italy, London, Routledge

Byé, M. 1928. Les lois des rendements non proportionnels, Paris, SireyChakravarty, S. 1987. Development Planning. The Planning Experience, Oxford, Clarendon PressChandler, A. D. 1977. The Visible Hand: The Managerial Revolution in American Business,

Cambridge, MA, Belknap PressChandler, A. D. 1990. Scale and Scope, Cambridge, MA, Belknap PressChandler, A. D. 2005. Shaping the Industrial Century: The Remarkable Story of the Evolution of the

Modern Chemical and Pharmaceutical Industries, Cambridge, MA, Harvard University PressChang, H.-J. 1994. The Political Economy of Industrial Policy, Basingstoke, MacmillanChang, H.-J. 2002. Kicking Away the Ladder: Development Strategy in Historical Perspective,

London, Anthem PressChang, H.-J., Andreoni, A. and Kuan, M. L. 2013. International Industrial Policy Experiences and

the Lessons for the UK, policy report for the UK Foresight Future of Manufacturing Project, UK Government Office of Science

Chenery, H. B. 1960. Patterns of industrial growth, American Economic Review, vol. 50, no. 4, 624–54

Chenery, H. B. and Taylor, L. 1968. Development patterns: among countries and over time, Review of Economics and Statistics, vol. 50, no. 4, 391–416

Coase, R. H. 1937. The nature of the firm, Economica, vol. 4, no. 16, 386–405Cole, W. A. and Deane, P. 1969. British Economic Growth 1688–1959, Cambridge, Cambridge

University PressEvans, P. 1996. Embedded Autonomy. States and Industrial Transformation, Princeton, NJ, Princeton

University PressFurtado, C. (1964) Development and Underdevelopment, Berkeley, University of California PressGeorgescu-Roegen, N. 1970. The economics of production, American Economic Review, vol. 60,

no. 2, 1–9Hall, P. A. and Soskice, D. (eds.). 2001. Varieties of Capitalism: The Institutional Foundations of

Comparative Advantage, Oxford, Oxford University PressHarcourt, G. C. 2006. The Structure of Post-Keynesian Economics: The Core Contributions of the

Pioneers, Cambridge, Cambridge University PressHarcourt, G. C. and Kerr, P. 2009. Joan Robinson, Houndmills, Basingstoke, Palgrave

MacmillanHicks, J. 1969. A Theory of Economic History, Oxford, Clarendon PressHirschman, A. 1958. The Strategy of Economic Development, New Haven, CT, Yale University

PressJacobides, M. G. and Winter, S. G. 2012. Capabilities: structure, agency and evolution,

Organisation Science, vol. 23, no. 5, 1365–81Kaldor, N. 1967. Strategic Factors in Economic Development, Ithaca, NY, State School of Industrial

and Labor Relations, Cornell UniversityKaldor, N. 1981. The role of increasing returns, technical progress and cumulative causation in

the theory of international trade and economic growth, Economie Appliquee, no. 4

at University of C

ambridge on A


Dow

nloaded from



Kuznets, S. 1971. Economic Growth of Nations: Total Output and Production Structure, Cambridge, MA, Harvard University Press

Landesmann, M. 1986. Conceptions of technology and the production process, in Baranzini, M. and Scazzieri, R. (eds.). Foundations of Economics, Oxford, Blackwell, 281–310

Landesmann, M. and Scazzieri, R. 1996. The production process: description and analysis, in Landesmann, M. and Scazzieri, R. (eds.), Production and Economic Dynamics, Cambridge, Cambridge University Press, 191–228

Landesmann, M. and Scazzieri, R. 1990. Specification of structure and economic dynam-ics, in Baranzini, M. and Scazzieri, R. (eds.), The Economic Theory of Structure and Change, Cambridge, Cambridge University Press, 95–121

Lazonick, W. 2007. Varieties of capitalism and innovative enterprise, Comparative Social Research, vol. 24, 21–69

Leon, P. 1967. Structural Change and Growth in Capitalism: A Set of Hypotheses, Baltimore, MD, Johns Hopkins University Press

List, F. 1885 [1841]. The National System of Political Economy, trans. Sampson S. Lloyd, London, Longmans, Green

Lowe, A. 1954. The classical theory of economic growth, Social Research, vol. 21, no. 14, 127–58Lowe, A. 1976. The Path of Economic Growth, Cambridge, Cambridge University PressMalthus, T. R. 1815. The Grounds of an Opinion on the Policy of Restricting the Importation of

Foreign Corn, London, John MurrayMarshall, A. 1920 [1919]. Industry and Trade, London, MacmillanMénard, C. 2004. The economics of hybrid organisations, Journal of Institutional and Theoretical

Economics, vol. 160, 345–67Morroni, M. 1992. Production Process and Technical Change, Cambridge, Cambridge University

PressMyrdal, G. 1958. Economic Theory and Underdeveloped Regions, London, MethuenNoble, D. F. 1984. Forces of Production: A Social History of Industrial Automation, New York, KnopfNurske, R. (1953) Problems of Capital Formation in Underdeveloped Countries, Oxford, Oxford

University PressO’Sullivan, E., Andreoni, A., Lopez-Gomez, C. and Gregory, M. 2013. What is New in the New

Industrial Policies?, Oxford Review of Economic Policy, forthcomingOcampo, J. A., Rada, C. and Taylor, L. 2009. Growth and Policy in Developing Countries.

A Structuralist Approach, New York, Columbia University PressPasinetti, L. L. 1981. Structural Change and Economic Growth: A Theoretical Essay on the Dynamics

of the Wealth of Nations, Cambridge, Cambridge University PressPasinetti, L. L. 1993. Structural Change and Economic Dynamics, Cambridge, Cambridge

University PressPasinetti, L.L. 2007. Keynes and the Cambridge Keynesians. A ‘Revolution in Economics’ to Be

Accomplished, Cambridge, Cambridge University PressPenrose, E. T. 1959. The Theory of the Growth of the Firm, Oxford, Oxford University PressPitelis, C. and Teece, D. 2009. The (new) nature and essence of the firm, European Management

Review, vol. 6, 5–15Powell, W. W. (1990) Neither market nor hierarchy: network forms of organization, in Research in

Organizational Behavior: An Annual Series of Analytical Essays and Critical Reviews, Greenwich, CT, JAI Press, 295–336

Quadrio Curzio, A. 1967. Rendita e distribuzione in un modello economico plurisettoriale, Milano, Giuffrè

Quadrio Curzio, A. 1975. Accumulazione del capitale e rendita, Bologna, Il MulinoQuadrio Curzio, A. 1986. Technological scarcity: an essay on production and structural change,

in Baranzini, M. and Scazzieri, R. (eds.), Foundations of Economics. Structures of Inquiry and Economic Theory, Oxford, Basil Blackwell, 311–38

Quadrio Curzio, A. and Pellizzari, F. 1999. Rent, Resources, Technologies, Berlin, SpringerRicardo, D. 1951 [1817]. On the Principles of Political Economy and Taxation, vol. I of Works and

Correspondence of David Ricardo, ed. P. Sraffa with M. H. Dobb, Cambridge, Cambridge University Press

Richardson, G. B. 1972. The organisation of industry, Economic Journal, vol. 82, no. 327, 883–86Robinson, J. 1956. The Accumulation of Capital, London, Macmillan

at University of C

ambridge on A


Dow

nloaded from



Rosenberg, N. (ed.). 1969. The American System of Manufactures: The Report of the Committee on the Machinery of the United States 1855, and the special reports of George Wallis and Joseph Whitworth 1854, Edinburgh, Edinburgh University Press

Rosenstein-Rodan, P. 1943. Problems of industrialization of eastern and south-eastern Europe, Economic Journal, vol. 53, 202–11

Sabel, C. and Zeitlin, J. (eds.). 1997. World of Possibilities: Flexibility and Mass Production in Western Industrialization, Cambridge, Cambridge University Press

Salais, R. and Storper, M. 1997. Worlds of Production: The Action Frameworks of the Economy, Cambridge, MA, Harvard University Press

Scazzieri, R. 1993. A Theory of Production. Tasks, Processes and Technical Practices, Oxford, Clarendon Press

Scazzieri, R. 2009. Structural economic dynamics: looking back and forging ahead, Economia Politica. Journal of Analytical and Institutional Economics, vol. 3, December, 531–57

Scazzieri, R. 2013. Increasing returns: towards a structural theory, Structural Change and Economic Dynamics, forthcoming

Simon, H. 1991. Organisations and markets, Journal of Economic Perspectives, vol. 5, no. 2, 25–44Smith, A. 1976 [1776]. An Inquiry into the Nature and Causes of the Wealth of Nations, gen. ed. R.

H. Campbell and A. S. Skinner; ed.W. B. Todd, Oxford, Clarendon PressSraffa, P. 1998 [1925]. On the relations between costs and quantity produced, in Pasinetti, L. L.

(ed.), Italian Economic Papers, vol. 3, Bologna: Il Mulino and Oxford University Press, 323–82Sraffa, P. 1960. Production of Commodities by Means of Commodities: Prelude to a Critique of

Economic Theory, Cambridge, Cambridge University PressSylos Labini, P. 1970. Problemi dello sviluppo economico, Bari, LaterzaSyrquin, M. 1988. Patterns of structural change, in Chenery, H. and Srinivasan, T. N. (eds.),

Handbook of Development Economics, Amsterdam, Elsevier, vol. 1, 203–73Wade, R. 1990. Governing the Market: Economic Theory and the Role of Government in East Asian

Industrialization, Princeton, NJ, Princeton University PressYoung, A. A. 1928. Increasing returns and economic progress, Economic Journal, vol. 38,

December, 527–42

at University of C

ambridge on A


Dow

nloaded from


triggers of change: structural trajectories and production dynamics

Documents