1

DIME Workshop on “Production theory” Process, Technology, and Organisation: Towards a useful Theory of Production

LEM | Scuola Superiore Sant'Anna, Pisa

A capability theory of production: learning in time, complementarities and proximities

Antonio Andreoni1 University of Cambridge

Abstract By integrating structural approaches to production with capability theories, the paper outlines a capability theory of production. The black box of the production process has been opened by describing it as a particular network of interrelated tasks through which transformations of materials are performed according to different patterns of capabilities coordination, and subject to certain scale and time conditions. Capabilities are structurally constrained, although an entire spectrum of virtual organizational and technological arrangements is possible. These possibilities are generated by multiple relationships of complementarity and similarity among structural components –i.e. materials in transformation, productive agents and network of tasks – both at the firm and inter-firm level. Discovering these possibilities is the very essence of a fully endogenous process of learning – i.e. structural learning in historical time. Complementarities are essential focusing devices in the continuous process of reconfiguration of the analytical map of production relationships. In this process of change, problems of capability compatibility and coordination arise both at the firm/industry and inter-firm/industry level. Different forms of production organization have historically emerged and reappeared in response to these coordination problems. Production organizations develop in historical time as a result of exogenous circumstances as well as endogenous dynamics according to patterns of specialization and diversification. In today’s global division of labour and knowledge, the need for coordinating closely complementary but dissimilar activities has found a response in the network-form of production organization. The effectiveness of these networks in coordinating production and innovation activities is determined by the degree of proximity among the agents embedded in the network and, thus, among their capabilities. As a matter of fact, the analysis of the internal structure of production and its change is essential for investigating the structural dynamics of economic systems and for developing a political economy of capability building. Keywords: Production Theory; Economics of Capabilities; Complementarities; Structural Learning; Network form of production organization; Proximity; Political Economy of Capability Building.

1 PhD Candidate, Department of Land Economy, University of Cambridge. Email: aa508@cam.ac.uk

2

Introduction Knowledge embodied in machines, human beings, institutional and organizational

structures is the most ‘powerful engine of production’ and the main driver of structural

change. However, the complex way in which knowledge enters production activating and

allowing processes of transformation of inputs into outputs is still not fully understood.

Cross-sector processes of knowledge diffusion, capabilities development through

learning in time are equally still not reconciled in a comprehensive theoretical

framework. Although many analytical efforts have been made, the ‘black box’ of

production still remains closed to economists limiting in many circumstances their ability

to propose effective policies for qualitative transformation of a country’s productive

structure – i.e. its development.

Today’s conventional production theory explains production processes as

relationships between combinations of productive factors – i.e. input quantities – and

certain quantities of outputs. By assuming that producers 'know-how' certain inputs may

be combined and transformed to obtain certain outputs, production functions do not make

any explicit reference to the capabilities needed to perform the real process of production.

Thus, in standard production theory there is no production process strictly speaking, and

knowing how to perform tasks and how to coordinate and activate different capabilities in

time, scale and among different agents are not problematic issues. The capability theory

of the firm, on the contrary, is based upon the recognition that the knowledge of

productive possibilities – i.e. input combinations – has to be complemented by the

availability of those capabilities – i.e. know-how – that are required to perform the

process of production. Although often neglected, this theory is at the very root of various

recent approaches based on different concepts of capabilities – i.e. productive,

technological and organizational – considered at different levels of aggregation – i.e.

firm, inter-firm, regional and national. By recognizing the complex nature of knowledge,

its tacit components as well as the complexities connected to its creation, diffusion,

absorption, adoption and accumulation, these contributions have shown theoretically and

3

empirically that the conventional understanding in economic theory of these dynamics is

very limited.

Starting from a critical analysis of these two different theoretical frameworks

(sections I and II), the paper argues that neither the production function approach nor the

capability theory of the firm alone are adequate in opening the black box of production as

well as in dealing with the emergent structures of the network form of production

organization. The main reason is that productive organizations of the network type

presuppose at the same time existing structural constraints, their evolution over time, and

a certain degree of open endedness as to what the actual organizational configuration will

be.

On this basis, the paper suggests that in order to reformulate the theory of

production in a 'useful' way we should focus on capabilities in a network-process

perspective. The analytical strategy suggested here is to complement the very common

functional focus characterizing capability approaches with a more definite structural

focus. Precisely, instead of concentrating the analysis in identifying capabilities

according to the functions and tasks they perform, the paper suggests a stylized

representation of the internal productive structure2 in which (i) the production process is

decomposed into elementary processes; (ii) interdependencies among structural

components –i.e. materials in transformation, productive agents and network of tasks –

are analysed; and (iii) structural change dynamics triggered by processes of learning in

historical time are given central stage. From a methodological standpoint, in order to

decompose the complex architecture of production as well as investigate learning

dynamics, the paper will maintain a separation between two fundamental level of

analysis3.

2 See Landesmann (1988) on the comparison between descriptive-analytical and optimizing approaches in production studies. 3 As suggested by Luigi Pasinetti in his ‘separation theorem’ (2007:255) ‘it is possible to disengage those investigations that concern the foundational bases of economic relations – to be detected at a strictly essential level of basic economic analysis – from those investigations that must be carried out at the level of the actual economic institutions, which at any time any economic system is landed with, or has chosen to adopt, or is trying to achieve”. See also Herbert Simon (1962) on techniques for decomposing the

4

Firstly, the paper opens the black box of the production process by describing it

as a particular network of interrelated tasks through which transformations of materials

are performed according to different patterns of capabilities coordination, provided

certain scale and time constraints (section III). At this level of the analysis the attention

will be directed towards detecting the different potential patterns of complementarities

among capabilities, tasks and materials, as well as among the virtual capabilities and

capabilities in use embedded in different fund factors. Not only time as well as scale do

play a crucial role in combining different capabilities for performing a network of tasks,

but also do affect the process of learning in the social organization of production.

Through a historical analysis of different processes of structural learning at the firm as

well as industry and inter-industry level, the paper will show how the process of learning

results from a continuous process of structural adjustment and discovery of

complementarities in historical time. As a matter of fact, the learning process is

cumulative, path dependent although reversible and in certain circumstances exclusive,

that is, the acquisition of certain capabilities might impede the acquisition of others of

different kind. Finally, it will be highlighted how learning takes place in time through a

relational process of intended collective action pervasively affected by serendipity

dynamics and unintended consequences.

At the second level of the analysis, problems of capability compatibility and

coordination are considered (both at the firm/industry and inter-firm/industry level) by

stressing the historical emergence and reappearance of different forms of production

organization. (section IV). As capabilities are distributed among different agents – i.e.

individual and collective funds, and must remain so if they want to develop according to

the principle of the division of labour, firms need not only to know how to do certain

things – i.e. direct capabilities, but also how to get other things done for them – i.e.

indirect capabilities. As for the latter, firms may be able to get things done by gaining

control of other capabilities (coordination by direction), by establishing continuous

architecture of complexity and Scazzieri (1993:11-13) on the distinction between social and technical division of labour.

5

relationships with others (coordination by inter-firm cooperation), or finally by market

transactions. Following the approach of G.B. Richardson, a coordination device will be

chosen simultaneously according to the degree of similarity and complementarities of

those activities that have to be performed. The historical analysis will suggest that the

emerging network form of production organization can be better identified as one of

producing in proximity. The paper concludes sketching out the kind of political economy

which would derive from a capability theory of production. A new political economy of

capability building would be focused on the following strategic issues: understanding the

structure of learning within the economy and across sectors; triggering processes of

knowledge flows, complementarities discovery and collective knowledge creation;

designing of ‘search networks’ for interacting agents; finally, identifying local industrial

policies for the creation/restoring of industrial commons.

I. Approaching production theories: in search for an analytical framework

The history of human societies is pervaded by extraordinary examples of the increasing

capabilities of individuals and collectivities, both as producers of wealth and consumers

of resources (Hicks, 1969; Rosenberg, 1976). Through continuous and cumulative

innovations, learning and processing of organizational and technological knowledge,

human beings have become increasingly capable of mastering their relationship with the

physical world. At the same time, the development of a sophisticated set of institutions –

i.e. social technologies – such as the market and the firm, has allowed human beings to

coordinate their social and economic relationships as well as to specialise in more

complex productive activities, both individually and collectively. In particular, the

adoption of the Smithian principle of the division of labour which allows shorter idle

periods of funds utilization, greater effectiveness in task execution and faster learning,

has been a fundamental step in the creation and improvement of specialist competences

because ‘knowledge grows by division’ (Loasby, 1999:50). Recognizing that structural

6

change dynamics are strictly related to the evolution of organizational and technological

capabilities as well as to their effective utilization in ‘supporting’ networks of productive

tasks of material transformations, economists have attempted many different analytical

representation of production processes (Scazzieri, 1993).

The analysis of the internal structure of production combined with a strong

attention for the representation of the complex system of interrelated production

processes were at the centre of the Classical theory of production. In particular classical

economists focused on the limited availability of non-producible goods, the utilization

problem and the various constraints determined by the production scale and its time

structure (Landesmann, 1988). Four are the main components of the Classical theoretical

framework. Francois Quesnay’s first formulation of the concept of productive

interdependencies called attention to the ‘circular flow’ of wealth production and

reproduction; Adam Smith’s analysis of the internal structure of the pin factory revealed

the microeconomic advantages of the division of labour – i.e. network of productive

funds – and the macroeconomic conditions on which it is based – i.e. stock of circulating

capital flows; Charles Babbage’s focus ‘on the causes and consequences of large

factories’ led to the formulation of the law of multiples and, thus, to the discovery of

different patterns of proportional utilization and maintenance of indivisible funds; finally,

Karl Marx’s analysis of different arrangements of production processes highlighted the

main features of the modern factory system and, thus, the working of the so called

‘collective machine’ (Landesmann, 1986; Scazzieri, 1993)4.

During the post-classical period, the production functions model emerged as the

main approach in production studies. In standard economic theory, production functions

represent complete sets of feasible input combinations for a given output; in an

isomorphic way, utility functions establish a relationship between combinations of

consumption goods and the satisfaction that they provide – i.e. utility. Both production

and utility functions are designed to show in the universe of rational choice and

4 The relevance of the classical ‘production principles’ for the development of a useful theory of production will be more widely considered in the following sections.

7

equilibrium allocations how the combinations chosen (respectively of inputs and

consumption goods) reflect relative prices. For conventional production theory does not

provide any analytical representation of the internal structure of production processes,

qualitative transformations generated by innovations and changes in the technology of

production remain completely unexplored. As a result, not only the network of tasks of

material transformations is relegated into a time-less black box, but also heroic

assumptions have to be made for what concern producers’ knowledge of the entire

spectrum of production possibilities as well as their availability of appropriate productive,

technological and organizational capabilities. On the contrary, as the literature on

localized technical change (Atkinson and Stiglitz, 1969; Antonelli 1995) has shown,

given the local and cumulative character of knowledge producers are only aware of a

limited number of factors composition laws – i.e. proximate production possibilities;

moreover, as shown in the capability literature, production ‘has to be undertaken by

human organizations embodying specifically appropriate experience and skills’

(Richardson, 1972:888)5.

A distinct theoretical framework in production studies stems from Wassily

Leontief ‘s (1947) pioneering input-output analysis and, more recently, from the fund-

flow model proposed by Nicholas Georgescu-Roegen (1969; 1970; 1971; 1986; 1990) on

which various contributions found their main inspiration (Scazzieri, 1983 and 1993;

Landesmann 1986; Morroni, 19926; Piacentini, 1995; Landesmann and Scazzieri 1996)7.

In the former contributions (as well as in Morroni, 1992) the focus is on the ‘different

patterns according to which the operations of fund factors are arranged over time’

(Scazzieri 1993:9); in the latter the productive process is represented as a particular

network of interrelated tasks through which a sequence of transformations of materials

5 It is beyond the scope of the present paper to provide a comprehensive discussion of the many contributions discussing the analytical and technical limitations of the production function models (see Robinson 1953-54; Pasinetti, 1959; Georgescu-Roegen 1969, 1970; Scazzieri, 1993; Sylos Labini 1995). 6 Morroni (1992, part III) is a unique attempt in operationalizing the fund-flow model of production. 7 See Mir-Artigues and Gonzales-Calvet (2007) for a review of these approaches.

8

are performed according to different patterns of capabilities coordination in quantity, time

and among productive agents.

Finally, a very influential attempt to cope with the fundamental limitations of

more conventional production models can be found in the capability theory of the firm,

an approach emerged at the intersection of various research areas, namely organizational

and innovation studies (Penrose 1959; Simon, 1969 Richardson 1972; Teece, 1980;

Langlois, 1992; Teece et al., 1997; Loasby 1999; Morroni, 2005), institutional and

evolutionary economics (Nelson and Winter, 1982; Lundvall, 1992; Dosi et al. 2000), as

well as empirical works in development economics (Bell 1982; Lall 1992; Bell and Pavitt

1995). On the one hand, given their specific (although quasi-exclusive) focus on the

capability dimension of production, these approaches are able to provide an in depth

analysis of the nature of the firm, the representation of different forms of knowledge,

important learning and technological dynamics (see section II); on the other hand,

however, capability approaches do not attempt to provide a comprehensive analytical

framework for understanding production processes, especially from the point of view of

the economics of structural change. Specifically, capability approaches find their limits in

their lack of understanding of production processes as systems of interdependent

processes, in the analysis of compositional changes and structural dynamics and, finally,

in detecting the structural constraints and patterns of complementarity by which

production and learning are pervasively affected (see section III).

The next sections will try to integrate structural approaches to production theory

with capability theories with the aim of outlining a capability theory of production. The

increasing interest for technological upgrading, learning, knowledge diffusion, structural

change and diversification make this attempt relevant. As a matter of fact, if we recognize

how development is mainly ‘a process that links micro learning dynamics, economy-wide

accumulation of technological capabilities and industrial development’ (Cimoli, Dosi and

Stiglitz 2009:543), increasing our understanding of production as a ‘network-process’

represents a crucial step both in a positive and normative perspective.

9

II. The Economics of Capabilities

In the Coasian theory of the firm (Coase, 1937) ‘production costs determine the technical

(substitution) choices [while] transaction costs determine which stages of the productive

process are assigned to the institution of the price system and which to the institution of

the firm’ (Langlois 1998: 186; see also Langlois 1992). Thus the firm emerges as the

more convenient way (lowest cost option) for obtaining control over the relevant cluster

of capabilities needed for undertaking the production process.

On the other hand, as theorized by Edith Penrose (1959), creating a firm may not

simply be a way of reducing transaction costs but it may be the highest value option for

the creation and development of capabilities8. Penrose's (1959:149) definition of the firm

as ‘a pool of resources the utilization of which is organized in an administrative

framework’ constitutes the original root of the capability theory of the firm. The firm is a

collection of physical and human resources which may be deployed in a variety of ways

to provide a variety of productive services. Refuting the conventional assimilation of

inputs and factors of production, Penrose introduced a fundamental distinction between

productive resources which are homogeneous and productive services which are

heterogeneous. In her interpretation of the growth of the firm, this distinction is

functional to recognize how ‘the services yielded by resources are a function of the way

in which they are used – exactly the same resource when used for different purposes or in

different ways and in combination with different types or amounts of other resources

provides a different service or set of services’ (Penrose 1959: 25)9. The growth process,

in the Penrosian framework, is realized through the recognition and exploitation by the

firm of ‘productive opportunities’, specifically of ‘all of the productive possibilities that

its entrepreneurs see and can take advantage of’ (Penrose, 1959:31). As Best (1999:108)

pointed out ‘productive opportunities links the firm to the customer in an interactive

relationship in which new product concepts are developed. The advances in productive 8 As Alfred Chandler's (1977) analysis of the growth of US firms demonstrates, an increase in transaction costs may be well justified if it allows an enhancement of capabilities and knowledge. 9 This intuition will be developed in the next section by discussing the issue of degree of utilization of the capabilities in use, provided certain funds of capabilities.

10

services can extend the firm’s productive opportunities by enlarging the members’

capacity to recognize and respond to new product concept possibilities in the

environment’10.

Penrose’s terminology (productive resources and services) was successfully

replaced by George B. Richardson (1972:888) who introduced to economics the term

capabilities. Maintaining the same distinction between resources and services,

Richardson describes industries and their firm as entities in which a large number of

activities are undertaken through the adoption of an appropriate cluster of capabilities.

‘It is convenient to think of industry as carrying out an indefinitely large number of activities, activities

related to the discovery and estimation of future wants, to research, development, and design, to the

execution and co-ordination of processes of physical transformation, the marketing of goods, and so on.

And we have to recognize that these activities have to be carried out by organizations with appropriate

capabilities, or, in other words, with appropriate knowledge, experience, and skills.’11

Capabilities are associated with clusters of relatively persistent ‘capacities to act’

(Cartwright, 1984). Thus they are know-how, both direct and indirect, that cannot be

deduced from (and reduced to) knowledge that. Moreover, since a capacity is not an act,

capabilities are particular dispositions to bring that act about as a result of intended

action. Capabilities may be characterized by different degrees of effectiveness and their

development is cumulative (but also path dependent) in the sense that ‘the acquisition of

certain kinds of know-how facilitates the acquisition of further knowledge of the same

kind, and impedes the acquisition of knowledge of incompatible kinds’ (Loasby

1999:58). The know-how emerge and accumulate through a continuous process of trial

and error, interpretations and falsifications, on the basis of an experimental and pragmatic

approach to the solutions of problems. However, the true fact that every human being

cannot be capable of performing alone all the activities necessary for the satisfaction of

10 Andreoni (2010) discusses the process of circular and cumulative causation through which consumer and producers capabilities co-evolve through historical time. 11 In contrast, Richardson (1972:887) critically asses the formal theory of the firm for being ‘little more than an application of the logic of choice to a particular set of problems’.

11

his/her needs implies that human societies have to solve a fundamental problem of the

coordination of capabilities and activities. In other words, we need not only to know how

to do certain things – i.e. direct capabilities – but also how to get other things done for us

– indirect capabilities. These latter capabilities are of two kinds: we may be able to get

things done by gaining control of other capabilities or by obtaining access to them12.

Capabilities are embedded in physical agents – i.e. machines and workers – as

well as organizational configurations and institutional arrangements. According to the

loci where they reside as well as the degree of aggregation considered – i.e. individual

agent, collective agent (e.g. organizations) or systemic (e.g. regional, national level) –

different capability concepts have been proposed. Many of them have been emerging

throughout the 1980s, putting the bases for a new Economics of Capabilities.

At the systemic level, Moses Abramovitz (1986) famously introduced the concept

of social capabilities to define those ‘tenacious societal characteristics’ that influence the

responses of people to economic opportunity. In the development of the catching up

hypothesis, Abramovitz equates social capabilities to managerial and technical

competences, but more crucially to a set of political, commercial, industrial, and financial

institutions. This systemic concept of capabilities have been also re-proposed in various

contributions on regional/national technological capabilities or innovation systems (see

for example Lall 1992; Lundvall, 1992).

At the individual level, a particular subset of human capabilities, that is, ‘the

ability to make independent technological choices, to adapt and improve upon chosen

techniques and products and eventually to generate new technology endogenously’

(Stewart 1981:80) were immediately recognized as fundamental factors in the expansion

12 As Marshall (1920) noted, evolution through the division of labour tends to favour both greater specialization (increasing capabilities) and closer integration (an increasing number of institutional devices to coordinate capabilities and activities). This idea was complemented by the famous aphorism by A.Young (1928) according to which ‘the division of labour depends upon the extent of the market, but the extent of the market depends upon the division of labour’. This means that ‘an increase in the market triggers further specialization which is a process that simultaneously increases the size of the market for specialist skills and activities’ (Best 1999:107). Thus, the division of labour is the fundamental premise for a process of division of knowledge and more effectively increasing capabilities. The development of new knowledge inspires a novel division of labour in a cumulative sequence that leads to the fundamental process of increasing returns (Marshall 1920; Young 1928).

12

of individual freedoms. On the one hand, Sen’s seminal work Capabilities and

Commodities (1985) focused on the link between the actual exercise of certain human

capabilities and the recognition of a set of individual entitlements13. What is central, here,

is that entitlements define the boundaries of the space of practicable capabilities, that is of

those productive/technical capacities people can exercise, given a certain socio-

institutional context and historical time. The need to identify the set of feasible operations

in production processes given a set of existing 'work capacities' or capabilities was also

initially stressed in the time-structure theory of production proposed by Landesmann

(1986) and Scazzieri (1993).

On the other hand, drawing from the Penrosian resource-based approach and on

previous contributions14, Bell and Pavitt (1995) propose the concept of technological

capabilities. The latter includes all those resources needed to generate and manage

technological change (including skills, knowledge and experience, and organizational

systems). Specifically, technological capabilities refer to the firm's abilities to undertake

in-house improvements across different technological functions, such as process and

production organization, products, equipment and investments. In the ‘technological

capability matrix’ proposed by Sanjaya Lall (1992:167; see below), firm-level

technological capabilities are categorized by technical functions (investment, production

and linkage capabilities) and their accumulation is actualized through the capacity of

performing more and more complex activities (from simple routine, to adaptive

duplicative activities, up to innovative risky activities)15.

13 A recent contribution (von Tunzelman and Wang, 2007) attempts a reformulation of Sen’s model of consumer capabilities for addressing the problem of capabilities utilization in production processes. Here, particular emphasis is given to the emergence of ‘dynamic interactive capabilities’. 14 Martin Bell (1982) distinguished two kinds of a firm’s fundamental resources: those needed to ‘operate’ existing production systems – i.e. productive capabilities – and those needed to ‘change’ production systems – i.e. technological capabilities. This contribution, together with the first empirical works developed by Sanjaya Lall (1982, 1987) and many scholars affiliated with the ‘Katz Programme’ in Latin America introduced the concept of technological capabilities. 15 The development of these capabilities is recognized to be both an outcome of an endogenous process of capabilities building within the firm and the response to exogenous stimuli, such as those arising from FDI made by transnational corporations into the global value chain (Archibugi and Pietrobelli 2003).

13

Those capabilities ‘needed to generate and manage technical change’ – i.e.

technological capabilities – very often ‘differ substantially from those needed to operate

existing technical systems’. The latter, that is, the ‘resources used to produce industrial

goods at given level of efficiency and given input combinations’ have been identified as

productive capabilities (Bell and Pavitt, 1995:78).

Table 1: Illustrative matrix of technological capabilities

Source: Lall (1992:167)

More recent contributions have refined and expanded the technological

capabilities approach by explicitly addressing the more organizational and managerial

dimensions of the firm. The first concept introduced is the one of dynamic capabilities,

that is, ‘firm’s ability to integrate, build and reconfigure internal and external

competencies to address rapidly changing environments’ (Teece et al., 1997: 516). This

set of capabilities is crucial in explaining differences in firms' competitive advantages, as

refers to the specific capacity of the firm to balance continuity – i.e. execution of

invariant processes – with change – i.e. transformation of capabilities, provided a certain

14

exogenous shock16. The second concept – i.e. organizational capabilities, arises from

evolutionary approaches which stem from the seminal work by Nelson and Winter An

Evolutionary Theory of Economic Change (1982, chapters 3-5). Here, as stressed by Foss

(2003), an attempt is made towards integrating the Simonian’s (1957) bounded rationality

notion and the Polanyi’s (1958) notion of tacit knowledge in a unifying framework.

Organizational capabilities account for a particular form of know-how, that is the

one which enables organizations to perform their ‘basic characteristic output actions –

particularly, the creation of a tangible product or the provision of a service, and the

development of new products and services’ (Dosi et al. 2000:1). Here, for organizations

‘to be capable of some thing is to have a generally reliable capacity to bring that thing

about as a result of intended action’ (Dosi et al., 2000:2). To the opposite of

organizational routines, which are characterized by a high degree of tacitness,

automaticity and repetitiveness, capabilities are developed and deployed by organizations

as a result of intentional and conscious decisions. However, as routines constitute one of

the building blocks of organizational capabilities as well as individual skills contribute to

the emergence of organizational routines, these two functional features of organizations –

i.e organizational capabilities and routines – remain strongly intertwined. Here, the

central point is to understand contextually to what extent a capability became routinized

and or a routine emerge as a distinct capability.

At this point, a general distinction is attempted: if technological capabilities

mainly refer to a form of know-how related to ‘how to handle the structure of nature’;

organizational capabilities specifically refer to the distinctive, collective and only

partially continuous ways to ‘handle the structure of the organization’. A theoretical as

well as empirical analysis of these different classes of capabilities is a fundamental step

in understanding how organizations develop, transform or disappear, learn and unlearn.

Although these approaches widely inform and potentially contribute to the construction

of a ‘useful’ theory of production, their main aim is to understand how single agents or

organizations, endowed with certain clusters of capabilities, can act as problem-solving

16 See also Kogut and Zander (1992) on the similar concept of ‘combinative capabilities’.

15

entities. In this sense they are characterized by a functional focus more than a truly

structural focus. The latter, however, is a crucial passage for understanding production as

a process. For example, those contributions in which more immaterial forms of

knowledge are stressed, do not fully recognize how production interdependencies and

material/scale/time constraints may both affect the utilization of capabilities and enable

the process of organizational coordination and/or reconfiguration of capabilities and task

combinatorics. Moreover the emergence of new forms of productive organizations

require a more detailed analysis of production processes which comprises and connects

micro-dynamics at the firm level, inter-intra industries interactions at the meso level and,

finally, division of labour and division of knowledge at the level of global production

networks. For these reasons, by adopting a ‘network-process’ perspective in which each

production organizations are described as ‘pools of fund complementarities’, the next

section will attempt to embed different concepts of capabilities in a structural theory of

production.

III. The production process: structural components, complementarities and learning in

historical time

Approaching production from the point of view of structural economics implies an

analytical focus on the following set of both quantitative and qualitatively coordination

problems: (i) how to synchronize the network of interrelated tasks; (ii) how to arrange

materials in transformation; (iii) how to organize and activate the network of productive

agents – i.e. funds of capabilities and flow agents17. Interdependencies among these

coordination problems are pervasive. Thus, the analytical space is one of structural

components in a multilayered network. As a second step (section IV), different forms of

17 Richardson (1972:885) stresses how ‘the habit of working with models which assume a fixed list of goods may have the unfortunate result of causing us to think of coordination merely in terms of the balancing of quantities of inputs and outputs and thus leave the need for qualitative coordination out of account.

16

production organization are introduced as structurally persistent coordination devices

(Landesmann and Scazzieri, 1996).

III.i Arranging structural components in time and scale: the network-process model A production process Pr (r = 1,…,k) can be represented as a particular network of

interrelated tasks through which a sequence of material transformations is performed

according to different patterns of capabilities coordination, provided certain scale and

time constraints.

Tasks refers to ‘what is sequentially and purposefully performed in a production

process’. Each task T j (j = 1, 2, …, J) can be decomposed in elementary operations or

clustered in groups of tasks.

Pr :

T

According to the set of capabilities and materials in transformation, tasks can be arranged

sequentially in various stages of fabrication (j = 1, 2, …, J), sometimes in a discrete way

others in a continuous way, that is, with or without interruptions. The latter distinction

reveals to be very relevant as soon as we consider how different forms of production

organization have historically developed different techniques for inventory and storage

capacities management (Marx, [1867]1972; Rosenberg 1994; Landesmann and Scazzieri

1996: chapter 8).

Materials refer to ‘what is transformed in the fabrication stages of a production

process’18. The relationship between materials and stages of fabrication can be

represented by a descriptive matrix M = [m ij ] in which any element refers to the material

i that has been transformed in the fabrication stage j.

18 In the case of ‘immaterial production’ – e.g. service activities – ‘materials in process cannot be identified, at least in the usual sense, and the production process generally takes the form of a close interaction among fund agents, in the course of which some of the characteristics of such agents (and sometimes their capabilities as well) may get transformed (Landesmann and Scazzieri, 1996:252-3).

T1 … Tj … TJ

…

17

M =

At each fabrication stage, only some materials will be utilized and, thus, transformed.

This implies that for each production process we will observe a certain realized matrix

M* = [m ij ] whose internal structure represents all the materials in use in the stages of that

production process (Scazzieri, 1999). As for the time structure, the material

transformation processes can be visualized as a system of pipelines (Landesmann, 1986;

see also Morroni 1992).

Productive agents refer to those ‘flow agents and fund agents utilized in

performing a network of tasks for material transformation’. The distinction proposed by

Georgescu Roegen (1970) between fund and flow inputs, calls attention to the degree of

permanence of a productive agent in the entire production process. Flow agents – e.g.

fuel, chemical catalysts, electricity, fertilizers – are utilized in certain stages of material

transformation; however, flow agents do not materially constitute the final output of the

process as the materials in use do. Flow agents used in a certain production process can

be described through a matrix F = [f ij ] in which any element refers to the flow agent i

that has been consumed in the fabrication stage j. As for materials, for each production

process we will observe a certain realized matrix F* = [f ij ].

F =

m11 … mij … mnJ

f11 … f ij … fmJ

18

On the contrary, fund agents are both mechanical artifacts – e.g. machines, tools,

equipments – or human beings – e.g. workers, supervisors, engineers, managers. Each of

them can be represented as a fund of multiple and complementary capabilities19. By

activating some of them, each productive agent is able to perform a single task or a set of

similar tasks, that is, tasks which requires the utilization of the same set of

complementary capabilities. In some cases, fund agents own capabilities which allow

only to execute productive tasks at a given level of technology – i.e. productive

capabilities; in others, fund agents present those sets of capabilities needed for improving

or even re-inventing the way in which tasks are performed – i.e. technological

capabilities. As discussed (see section II), performing a production process requires the

coordination of different bundles of productive and technological capabilities embedded

in multiple fund agents. Here, the concept of organizational capabilities comes in to

identify the specific capacity of a production organization to coordinate different

productive and technological capabilities embedded in different fund agents. This

specific set of capabilities points to ‘the importance of the immaterial side of production,

that is, of the complex network of cognitive rules and practices, customs and social norms

from which production is made possible’ (Landesmann and Scazzieri, 1996:4)

As different production organizations are endowed with different organizational

capabilities, they can arrange the network of interrelated tasks of material transformation

– i.e. their cognitive map – in different ways. In other words, multiple cognitive maps of

production are possible given the fact that the same process can be performed according

to different tasks arrangements, by coordinating different fund agents, by utilizing

different sets of capabilities or, finally, by transforming different materials. This point

suggests that, the production process observed in a certain production organization is just

one realization of an entire spectrum of production possibilities, satisfied certain time and

scale constraints (see section III.iii).

19 Fund agents maintain their characteristics substantially unaltered during a production process, provided certain tolerance thresholds are not violated (Landesmann, 1986).

19

Having that in mind, we can distinguish between the bundle of capabilities owned

by a certain number of fund agents in a given productive organization – i.e. virtual

capabilities – and the capabilities utilized by the same productive organization in

performing a certain network of tasks – i.e. capabilities in use20. A way to represent the

bundle of capabilities and capabilities in use is to consider a matrix C = [cij ] in which any

element cij denotes the relationship between the capability i and the task T j performed at

the stage of fabrication j. As shown in Cartwright (1989), very often capabilities can be

expressed in a quantitative form, so we can assume that they are comparable in cardinal

space (Landesmann and Scazzieri 1996:197)21. As for the matrix of materials M* = [m ij ],

the internal structure of the matrix C* = [cij ] shows among the capabilities (vis-à-vis

materials) available those which have been in use in performing each task. A series of

specifications are needed.

C =

First of all, the distinction between the matrix of virtual capabilities C = [cij ] and

the one of capabilities in use C* = [cij ] stresses how the same production process Pr (r =

1,…,k) can be performed by using different bundles of fund agents, that is, different

combinations of capabilities. In other words, given a virtual capability space C = [cij ], the

production process Pr can be represented with different matrices C1* = [cij ] ≠ C2* = [cij ]

≠ … according to the different fund agents in use. However, even when two different

productive organizations perform the same process Pr by combining the same bundle of

20 As it has been stressed, this distinction leads to interpret the emergence of new productive structures within the space of virtual practices as ‘the outcome of a clustering process that brings about a rearrangement of the primitive elements of productive activity; [thus] structural change may be considered as a case of variation within a spectrum of virtual possibilities’ (Scazzieri 1999:230) 21 For a review of recent attempts in measuring technological capabilities at the country level see Archibugi and Coco, 2004.

c11 … cij … cqr

20

capabilities C* = [cij ], these capabilities can be employed in different proportions. For

example, two firms F1 and F2 can perform the same production process by using the same

two fund agents – i.e. workers w and machines m – but one of them is labour intensive

(the ratio w/m > 1) while the second one is capital intensive (the ratio w/m < 1).

C1*=

5w 7w 1m 1m

Thus, by comparing C1* with C2* we can discover specific features of the production

process Pr performed by F1 and F2 with different proportions of the same bundle of

capabilities. In particular, the two matrices express different relationships of

complementarity among funds of capabilities. In our case, the first stage T1 of the

production process Pr can be performed either by combining one machine with five

workers or three machines and one worker (see above). Given these relationships of

complementarity, the kind of combinations of fund agents that firm can select from the

virtual capabilities space C = [cij ] for performing Pr are limited. Moreover, scaling up the

production process not only requires the consideration of these relationships of

complementarity but also that a law of proportionality among all the structural

components of the process is satisfied (see below).

A second specification is related to the time structure of the production process in

relation to the capabilities in use. The matrix C* = [cij ] can be transformed in a matrix of

capabilities use – times Ω* = [ Ωij ] where the generic Ωij represents the use-time of the

capability cij in the production process Pr.

Ω* =

Ω11 … Ωij … Ωqr

C2* =

1w 2w 3m 3m

21

Given the case in which two different productive organizations perform the same process

Pr by combining the same bundle of capabilities C* = [c ij ], by comparing the matrices of

capabilities use – times Ω* we can discover if one time arrangement of Pr is more or less

time wasting than another. For example, the reconfiguration of the time structure of Pr

from one in line to one in parallel can reduce the amount of time waste of fund agents

across fabrication stages (see below).

Finally, the analysis of the matrix C = [cij ] suggests how a production process

may be qualitatively transformed even without equipping the productive organization

with new bundle of capabilities, but just re-arranging capabilities among the network of

tasks which have to be performed. Moreover, as a matter of fact, there are various way of

combining elementary operations into tasks or clustering tasks. Evidently, this depends

again on the capabilities embedded in funds agents, their degree of utilization, their

distribution among productive agents, their time arrangements as well as the kind of

materials in transformation utilized (see section III).

The fact that production processes are characterized by multiple

interdependencies among structural components (Buenstorf, 2005), can be visualized by

mapping in Venn diagrams the relationships between capabilities, tasks and materials22.

Consider the example of the production process Pr (r = 1,…,k) for the commodity A,

given the matrices C = [cij ] M = [mij ] and F = [f ij ]. Venn diagrams represent

respectively the spaces of capabilities C, tasks T and materials M available in a given

organization which produces the commodity A. Given the entire spectrum of possible

combinatorics contained in Venn diagrams, we can visualize the analytical map of

production relationships among C, T and M (see figure 1)23. The mapping from the

capability space C to the task space T – i.e. job specification programme – can be

determined following different criteria. For example different capabilities may be

22 For clarity, the flow agents are taken out of the picture. The choice to privilege the other three dimensions matrices C, T and M is related to the fact that commonly there are higher degrees of freedom in their combinatorics and the use of flow agents is strictly dependent on the utilization of fund agents. 23 The concept of ‘analytical map of the true [interpersonal] relations’ is proposed in Georgescu-Roegen (1976:205) as one possible realization of the ‘entire spectrum of peasant institutions’. See also Baranzini and Scazzieri (1990:252-255).

22

relatively more or less adequate for the execution of one task or cluster of tasks; in other

cases a reconfiguration of the job specification programme – i.e. different mapping C

T – allow to activate some capabilities or to achieve higher efficiency in the utilization of

capabilities in use.

C M

T

C = [cij ] Pr : M = [mij ] t

Figure 1: The analytical map of production relationships

Source: author

The network of relationships and interdependencies among the spaces C, T, M has

to be synchronized over time and according to specific scale requirements determined by

the existence of process indivisibilities as well as indivisible productive agents24. As for

the time structure, synchronization has to be pursued at three different levels – i.e.

coordination and utilization of fund capabilities, arrangement of interdependent tasks

over time, transformation of materials over time. The difficulty of matching the ‘time

sequencing requirements’ of these three dimensions makes perfect synchronization

24 For a comprehensive discussion of time and scale as structural dimensions of production see Landesmann, 1986; Morroni 1992; Scazzieri 1993; Landesmann and Scazzieri 1996.

C11

C23

C32

Cij

C41

T3

T2

T1

m11

m21

m32

T1 … Tj … TJ

…

23

impossible as well as justifies the co-existence of patterns of simultaneity or

sequentiality, for example in the utilization of capabilities funds. Thus, time gaps and

wastes are structurally determined and only partially reducible (Landesmann and

Scazzieri, 1996).

As for indivisibilities, processes are indivisible when they are not ‘indifferent to

size’. In biology as well as technology all individual processes ‘follow exactly the same

pattern: beyond a certain scale some collapse, others explode, or melt, or freeze. In a

word, they cease to work at all. Below another scale, they do not even exist’ (Georgescu

Roegen, 1976:288). The fact that processes are ‘scale-specific’, that is, are characterized

by upper and lower bounds, implies that activating smaller or larger scales of the same

process can be done only if a law of proportionality among the structural components of

the process is satisfied. In other words, a law of multiples must be satisfied (Babbage,

1832). At the level of the structural components, limitations both in the bundling and

unbundling of capabilities or in the size of flow agents – i.e. productive agents

indivisibilities – as well as in the decomposability of materials in transformation are

widely present. However, flow agents are more often divisible than funds of capabilities

such as ovens, machines, equipment, tools, containers.

The existence of indivisible funds of capabilities fundamentally arise from the fact

that for being a fund of capabilities fully utilized a specific scale of production of a

certain commodity A has to be achieved. Thus, the existence of invisibilities implies that

for lower scales of production of the same commodity A, funds of capabilities would be

underutilized. However, if funds of capabilities are not too specialized in the execution of

some tasks, scale constraints can be overcome by utilizing indivisible funds of

capabilities in producing other commodities B, C... . Processes of production of these

commodities has to be enough similar to the one executed for producing A, which means

they require the same bundle of capabilities. This example suggests how the job

programme C T as well as the internal structure of production organizations

historically emerged are very much scale dependent.

24

In sum, ‘constraints with respect to the time structuring and with respect to scale

are the two fundamental constraints on the internal organizational structure of a process

and these constraints are of course a function of the state of technological (or process)

knowledge’ (Landesmann and Scazzieri, 1996:206).

III.ii Complementarities and similarities: structural features of production processes Complementarities as well as similarities among structural components of a production

process are pervasive and can realize at the firm as well as at the inter-fim level. In

particular, funds of capabilities as well as ‘innovations hardly ever function in isolation’

(Rosenberg, 1979:26). Instead, their utilization and productivity critically depend on the

availability of complementary technologies or, more generally, of other productive agents

or materials25.

At the firm level, each fund agent involved in a production process can be

assigned to perform some elementary operations, one single task –i.e mono task – or a

cluster of similar tasks – i.e. multi task. In general, when the same fund agent executes

more than one task, these tasks must be not too dissimilar as fund agents are endowed

with only a limited set of complementary capabilities, exactly those which allows them to

perform that limited set of similar tasks. As a matter of fact, if a production process is a

network of similar tasks, fund agents can be arranged and utilized in many different ways

in the execution of tasks. Thus, in this case, the arrangement of tasks in the production

organization can be very flexible and each fund agent can be involved sequentially in all

the tasks of the process – i.e. high degree of fund agents substitutability. On the contrary,

if a production process is constituted by a network of dissimilar tasks, fund agents will

not be able to perform all the tasks. A production process composed by a network of

dissimilar tasks will be executed by specialized fund agents – i.e. low level of fund agents

substitutability. In other words, if tasks are very dissimilar and complex, fund agents have

to specialize in the execution of only one task, or even in performing elementary

25 See Rosenberg (1979:28) for an analysis of how ‘the social payoff of an innovation can rarely be identified in isolation’.

25

operations of most complex tasks. In this case, a number of processes of the same type

can be organized in series so that specialized fund agents can be utilized in performing

the task in which they are specialized without long time of inactivity. This arrangement

allows firms to reduce wastes of time as fund agents will shift over time from one

production process to another one.

In a production process constituted by a network of interdependent dissimilar

tasks, capabilities funds performing a specific task in one stage of fabrication are

combined with others performing other tasks in other stages of fabrication with a

relationship of complementarity rather than of substitutability. Of course, this

relationships of complementarity can be also present among the set of capabilities funds

which execute the same complex task at one stage of fabrication. Both inter-tasks

complementarities and intra task-complementarities are scale-dependent – e.g. at the intra

task level, each machine requires a fixed number α of workers and technicians to execute

a certain task. Thus, indivisibilities accounts for many forms of complementarity in

production processes. Finally, complementarities may also arise among different

production processes when the combined execution of tasks reduced the costs of each

production process – i.e. economies of scope. The reduction of costs usually depends

from a better coordination and utilization of indivisible capabilities funds. To

recapitulate, at the firm level, complementarities arise among indivisible fund agents,

intermediate stages of fabrication of a single production process, combination of various

production processes (Morroni, 1992; 2005).

At the inter-firm level, the classical paper ‘The Organization of Industry’ by G.B.

Richardson (1972) shows that various forms of inter-firm cooperation we observe in

reality arise from different patterns of similarity and complementarity among firm’s

activities. In the Richardsonian framework, the production of each final commodity is

broken down into various stages or activities, each of them executable by various firms.

‘Activities which require the same capability for their undertaking’ are called similar

activities (Richardson 1972:888). On the other hand, activities can be complementary

‘when they represent different phases of a process of production and require in some way

26

or another to be coordinated [...] both quantitatively and qualitatively’ (idem :889-890).

Coordination can be realized in three different ways by direction 'when the activities are

subject to a single control and fitted into one coherent plan' (see above), by cooperation

when 'two or more independent organizations agree to match their related plans in

advance' or finally through market transactions. At this stage of the analysis, where only

structural elements have been taken into consideration, we limit ourselves in stressing

how according to Richardson (1972:892) the main explanation behind complex and

interlocking clusters, groups and alliances of firms can be found in the need to coordinate

‘closely complementary but dissimilar activities’26. A typical example is the one of two

firms owning different technological capabilities which decide to cooperate for

performing dissimilar but closely complementary activities. This example leads towards a

fundamental intuition. On the one hand, as complementarities introduce qualitative and

quantitative interdependences both at the firm and inter-firms levels, they can be

considered as constraints. However, on the other hand, as complementarities

continuously force firms in re-configuring their production processes, in re-imagining

their organizational form, in developing new capabilities, in expanding their external

network, they can be considered as focusing devices for innovation. Thinking of

complementarities not only as constraints but also as focusing devices is a perspective

widely supported by the historical analysis of processes of structural learning.

III.iii Structural learning: complementarities as focusing devices The main trait of a capability theory of production, is that capabilities are structurally

constrained. At the same time, a pervasive element of virtuality characterizes this

framework. The virtual component has been introduced suggesting how the coordination

problems in the space of productive agents, materials and tasks can be solved in multiple,

although interdependent, ways. In other words, as stressed by Salais and Storper (1997),

26 The emergence of different forms of production organization, patterns of complementarities and diversification are analysed in section IV.

27

there are ‘worlds of production’ – i.e. a variety of production programmes. Thus, ‘worlds

of possibilities’ are open for transforming production and its outcome – i.e. process and

product innovations (Sabel and Zeitlin, 1997). This statement does not want to

underestimate the fact that these possibilities – i.e feasible organizational and

technological arrangements – have to be known to be exploited and that the existence of

indivisibilities, bottlenecks, technical imbalances, complementarities, materials

characteristics are pervasive constraints. On the contrary, it does stress how discovering

these possibilities, given certain structural constraints, is the very essence of a fully

endogenous process of learning. Complementarities, in particular, have historically

resulted in being crucial focusing devices in the process of choice and exploration of new

techniques (Sen, 1960; Rosenberg, 1969, 1976, 1979, 1982; Richardson, 1972).

Recognizing complementarities as focusing devices means to investigate how

'[c]omplex technologies create internal compulsions and pressures which, in turn,

initiate exploratory activity in particular directions' (Rosenberg, 1969:4). As shown by

Nathan Rosenberg’s work, historical episodes of technological change are invaluable

heuristics in identifying those ‘inducement mechanisms’ (Hirschman, 1958), ‘compulsive

sequences’ and causational chains through which a process of structural learning in

historical time realizes. Rosenberg (1969) identifies three main inducement mechanisms,

namely technical imbalances or bottlenecks, labour-saving/uncertainty-reducing

machines, substitutes or alternative sources of supply. Few examples can help at this

point.

In 1900 the machine tool industry was revolutionized by the introduction of high-

speed steel which allowed increasing the hardness of cutting tools. However, ‘it was

impossible to take advantage of higher cutting speeds with machine tools designed for the

older carbon steel cutting tools because they could not withstand the stresses and strains

or provide sufficiently high speeds in the other components of the machine tool’

(Rosenberg, 1969:7). As a consequence, structural transmissions, control elements and

other machine tool components had to be redesigned; this change ‘in turn, enlarged

considerably the scope of their practical operations and facilitated their introduction into

28

new uses’ (Rosenberg, 1969:8). This typical example of a technical imbalance leading to

changes in complementary processes as well as structural components, highlights how a

technical constraint can actually activate a process of exploration and searching in which

‘the size of the discovery need bear no systematic relationship to the size of the initial

stimulus’.

Not only technical, but also social processes can work as triggers for

technological change. In the Poverty of Philosophy Karl Marx observed how ‘after each

new strike of any importance, there appeared a new machine’ (n.d.:134; first source

Rosenberg, 1969). The threat of strikes introduces a critical element of uncertainty in the

supply of labour and strongly affects the delicate time structure of a production process.

Robert’s self-acting mule and the Jacquard punching machine in the nineteenth century;

the introduction by the British Government of the ‘American System of Manufacturing’

in the gun making industry in 1854; even the introduction of legal restrictions, are all

cases in which the invention or acquisition of a new machine is just the first step of a

subsequent endogenous process of structural learning (Rosenberg, 1969; see also Chang,

2002). For example, becoming a new machine available, productions which were

technically feasible but not economically convenient become possible. This possibility

may depend on increasing the scale of complementary machines or in the re-arrangement

of workers in the production unit. Moreover, as soon as the scale of production increases

‘a shifting succession of bottlenecks’ will emerge. Focusing on them, engineers will start

exploring new possible structural configuration of the production process, a tour which

may lead to discovering serendipitously ‘singleton techniques’ (Mokyr, 2002)27.

One of the possible ways to visualize the dynamics behind these historical

examples is to make use of the analytical map of production relationships28 proposed

27 Other examples can be found in analysing how complementary innovations such as refrigerators and railroads or steamships have affected the reduction of transportation costs, increased the degree of regional specialization, allowed increasing returns to scale. 28 These dynamics may be also analysed by adopting the virtual matrix of capabilities C = [cij ] and the matrix of capabilities in use C* = [cij ] (see above). Here, through the process of structural learning some relationships of complementarity among funds of capabilities will end, others will change although maintaining the same proportions among fund agents, others will be completely transformed.

29

above and to show how structural learning realize in two steps: (i) identifications of focal

points – e.g. technical imbalances, bottlenecks, scale constraints, material innovation,

new output requirements – and (ii) discover of new complementarities which act as

focusing devices for gradual improvements or radical innovations. Different examples

can be visualized in the analytical map of production relationships (see figure 2). Let’s

consider the following illustrative case. The acquisition of a new fund of capabilities c53 –

e.g. a new machine through technology transfer or a traditional machine transformed by

small improvements – can trigger a cascade process of production reconfiguration. The

task T2 has to be decomposed in the similar tasks T’2 and T’3; the obsolete fund of

capabilities c23 can be dismissed; the discovery of a new material m54 requires to execute

a new task T4 to be transformed; the scale of the capability c41 given the introduction of a

new indivisible fund of capabilities c53 has to be changed; a previously unutilized

capabilities fund c64 is activated as a new complementarity with the new material m54 has

been discovered.

C = [cij ] Pr: M = [mij ] C’ = [cij ] P’r: M’ = [mij ]

Figure 2: Structural learning: complementarities as focusing devices

Source: author

m11

m21

m32

C11

C23

C41

C64

C32

T1

T3

T2

T’ 2

T’ 3

T4

C53

m43

m54

T1 T2 T3

T1 T’ 2 T’ 3 T4

30

However, this cascade process of production reconfiguration cannot be thought as

an automatic one, as Rosenberg himself recognizes. In order to react to structural

feedback firms have to be equipped with a certain bundle of organizational and

technological capabilities. This point has been raised and historically documented in

Poni’s analysis of the silk industry. From his comparative study of the silk industry in

Lyon and Bologna we realize that, in order to activate a process of structural learning,

feedback must ‘end in the hands of the right persons [as feedback management] require

capabilities and knowledge of techniques which are not necessarily available in the right

moment, in the right sector, in the right hands’ (Poni, 2009:297; author’s translation). The

same condition has to be satisfied in the case of learning by using, a process which

occurs ‘only after certain new product are used’ (Rosenberg, 1982:122). Thus, as stressed

before, same technical imbalances can be either constraints or focusing devices for new

opportunities discovery according to the capacity of the productive organizations to see

and take advantage of them (Penrose, 1959:31).

In sum, what the concept of structural learning in historical time is aimed to

stress is that a continuous process of reconfiguration of the analytical map of production

relationships occurs at the firm and inter-firm level and that complementarities are

focusing devices for technological change. However, given the multilayered nature of

production, in complex production organizations multiple actors experience at the same

time various processes of structural learning in different directions and at different

speeds. For this reason, two aspects have to receive more attention, respectively the time-

speed and the collective dimensions of learning.

The first dimension is still widely unexplored. Learning in time can proceed at

different speeds according to the time required for reconfiguring the structure of

production, given a certain exogenous or endogenous factor, or according to the time

knowledge requires to flow – i.e. being disseminated and adsorbed – throughout the

production organization or at the inter-firm level. In other words, the problem is not only

‘how to learn to learn’ or ‘what to learn’ but also ‘how to learn faster’. As shown by

Dodgson (1991) the differential ability in learning quickly about technological

31

opportunities is a crucial determinant especially in those sectors – e.g. biotechnology –

characterized by an uncertain and generally fast process of transformation.

As for the second dimension, Lundvall (1992) introduced the concept of learning

by interacting as a critical feature of societies. Capabilities are collectively developed

through social interactions mainly by observing and imitating others’ actions as well as

by mirroring their attitudes. The organizational design as well as the underlying relational

structures can affect people’s disposition towards mutual learning and knowledge

discovery which can results from the same process of learning – e.g. small improvements,

process and product adjustments, etc. In the analysis of the bounded rationality problem,

Herbert Simon (1957) introduced the idea that individuals’ learning is socially

constructed or, in other words, that ‘[w]hat an individual learns in an organization is very

much dependent on what is already known to (or believe by) other members of the

organization and what kinds of information are present in the organizational

environment’ (Simon, 1991:125). Historically, learning by interacting realizes in various

ways from more co-operative to more competitive forms such as copying, foreign skilled

workers’ recruitment, technicians exchange, pooling of technology, organization of expos

and industrial espionage (Chang 2002; Poni 2009). One of the main drivers of today’s

network forms of production organization has to be found exactly in the many

opportunities of structural learning it provides.

IV. Forms of production organization: ‘producing in proximity’ through networks

Production processes and structural learning in historical time require the operation of a

complex organizational structure. The notion of forms of production organization

captures the different ways in which ‘coordination problems have been resolved in

particular circumstances, taking into account the state of technological knowledge, the

evolution of patterns of demand, natural resources and environmental constraints, etc.’

(Landesmann and Scazzieri, 1996:218). The emergence and disappearance of different

32

forms of production organization testify that the coordination of tasks, productive agents

and materials in transformation can follow different patterns according to specific

objectives and constraints (see above)29. Thus, the ‘virtual coordination patterns’

actualise as ‘real responses’ to specific historical and contextual circumstances. For

example, the job-shop model, adopted in the craft system, is a form of production

organization characterized by (i) multi-task productive agents performing interlocking

networks of similar tasks and (ii) a ‘stop and go’ process of material transformations.

These two features provide craft system with high flexibility and adaptability in solving

unexpected problems, although low capacity in satisfying increasing levels of demand.

The transition from the job-shop model to the putting-out or the factory forms is a

response to the considerable growth of demand for certain commodities and emerging

technical constraints. For example, the increasing complexity of tasks or the need for

higher speed at which material transformations follow one another30.

The differentiation of tasks and the increasing specialization of capabilities – i.e.

from multi- to mono- tasks – are the common features of the putting-out and the factory

form of manufacturing31. However, these models developed ‘different ways of dealing

with the interface between different stages of any given transformation process’

(Landesmann and Scazzieri, 1996:262). On the one hand, the putting-out model (also

known as Verlagssystem) is structured as a network of separate ‘specialized workshops’,

each of them performing a limited number of tasks related to a specific stage of

fabrication – i.e. craft differentiation. Very often the workshop (or the merchant)

executing the final stage of production is responsible for the coordination of the flexible

29 As stressed by Pasinetti (2007:271) ‘[The production paradigm cannot] abstract, as the models of exchange usually do, from historical specificities, since the kind of institutions that shape an industrial society, besides being far more complex, are inherently subject to changes induced by the evolving historical events, much more extensively than those that shaped the era of trade’. 30 For a detailed analysis of these forms of production organization see Scazzieri (1993) and Landesmann and Scazzieri (1996: chapter 8). 31 The analysis of different forms of production organization focuses on manufacturing activities, defined as processes of transformation of raw materials into useful products. However, it is possible to conduct the same analysis taking into consideration other sectors. For example, see Romagnoli’s essay (1996) in Landesmann and Scazzieri (1996) for an analysis of forms of production organization in agriculture.

33

linkages among network’s nodes. Sometimes, it is also involved in previous stages of

fabrication, for example by assuring the provision of raw materials (Hicks, 1969). On

the other hand, the factory model was developed as a concentrated form of production in

which complex tasks were subdivided in elementary operations performed by highly

specialized productive agents. Here, both workers and machines – i.e. indivisible funds of

capabilities – are coordinated in a way that guarantee their full and continuous utilization

in executing networks of dissimilar tasks (Landesmann, 1986:294). Moreover, thanks to

the precise coordination of tasks and materials in transformation over time, the factory

model allows to respond to customers’ request ‘just in time’. However, a just-in-time

network may consist of a mix of factory establishments and job-shop workshops (see

Scazzieri, 1993:90). The continuous improvements in the factory form of manufacturing

led to the development of the flexible manufacturing system as well as the achievement

of increasing coordination of materials in transformation over time.

The historical analysis of different forms of production organization permits to

identify among others two main stylized facts. First of all, forms of production

organization structurally persist for a certain period of time and, very often, old ones

compete with new ones for considerable periods in history. The reason is that transition is

costly and require adaptation and reconfiguration of production through structural

learning (Landesmann, 1988:173; Rosenberg, 1976). Moreover, as each form of

production organization is a response to specific socio-technical circumstances, old ones

can reappear. Examples are today’s re-emergence of ‘modern craftsmen’ (Sennett, 2008)

adopting the job-shop form of production organization32 as well as the increasing

pervasiveness of global production networks, a model which share many features with

the putting-out system developed in early modern Europe (Dyker and von Tunzelman,

2002). Secondly, historical examples show that production can be coordinated by

adopting various devices, namely (i) by market transactions; (ii) by directing and

controlling production activities within a single organization; or (iii) by collaborative

32 See also Andreoni and Pelligra (2009) for an analysis of various forms of relational credit adopted for financing modern micro-enterprises and craftsmen.

34

arrangements of varying degrees of intensity, between organizations which are formally

independent.

The classification of production activities along the dimensions of similarity and

complementary proposed by Richardson (1960; 1972) contributes in explaining (i)

patterns of diversification within a single organization, and (ii) patterns of specialization

in interlocking networks of production organizations. Diversification in similar

production activities do not require any investment in building or acquiring new

capabilities. Instead it allows to fully exploit firm’s technological and organizational

capabilities. For example, indivisible funds can be more efficiently utilized and, thus,

both economies of scale and scope achieved. In this case, activities will be directed and

coordinated within the firm which has an incentive in diversifying in those similar

activities for which it owns all the necessary capabilities. On the contrary, diversification

in closely complementary but dissimilar activities, such as producing ‘a particular car

with a particular brake and a particular brake lining’ in the same firm, require

investments for building or acquiring new bundles of capabilities and, very often, a

complete reconfiguration of the job programme C T. In this second case, then,

coordination ‘cannot be left entirely to direction within firms because the activities are

dissimilar, and cannot be left to market forces in that it requires […] the matching, both

qualitative and quantitative, of individual enterprise plans’ (Richardson, 1972:891-892).

The coordination device adopted in the case of closely complementary but dissimilar

activities will be the network form of production organization33. This network will be

composed by independent and highly specialized production organizations, each of them

performing in a coordinated way a set of closely complementary but dissimilar

activities34.

Relationships among productive nodes in producers’ networks generally realize in

multiple dimensions – i.e. not only capabilities, but also materials and flow agents – as

33 See Powell (1990) for a classical analysis of network forms of organization. 34 With this respect Silver (1984) argues that ‘in developing countries, or in developed economies when innovation renders the market’s existing capabilities obsolete, a firm may have to integrate into many dissimilar activities in order to generate all the complementary activities it needs’ (Langlois, 1992:108)

35

well as at multiple levels – i.e. regional, national or global. The dimension of connection

among two (or more) nodes in a network is not irrelevant – e.g. being part of a capability

vis-à-vis a material network. Actually, it introduces an important qualitative distinction.

The fact that firms in a region are strongly interconnected with global producers does not

guarantee that they are also networking in the most strategic dimensions. A capability

theory of production stresses how the most relevant dimensions are those related to

capabilities. As a matter of fact, ‘it is knowledge stocks within firms and knowledge

flows to them, between them and within them which underlie change in the types of

goods they produce and the methods they use to produce them’ (Bell and Albu,

1999:1722). In his Principles of Economics, Alfred Marshall (1920) suggested that every

firm builds up an external organization as a means not only for developing a special

market and establishing preferential relationships with costumers, but also for acquiring

that kind of knowledge that cannot be attained by anonymous contracting35. In the

network form of production, the transfer and pooling of technology, drawings, tools,

personnel are the main triggers of structural learning dynamics and, thus, the most

strategic dimensions of connection36.

However, the effective functioning of network forms of production organization

is very much determined by the degree of proximity among the different production

organizations embedded in the network (Bellet et al., 1993; Boschma, 2005). Here, the

concept of proximity should be declined in a plural way in order to stress its

multidimensional and multilevel nature.

As for the former aspect – i.e multidimensionality – proximity does not refer only

to a condition of geographical distance. In fact, geographical proximity is neither a

necessary nor a sufficient condition; in some circumstances it can be even

counterproductive by discouraging geographical openness and the activation of 35 The creation of these continuing relationships is based not only on the establishment of forms of cooperation between different firms, as well as among customers and firms, but also on the same competitive dynamics that are achieved through market transactions. As Richardson (1972:896) eloquently states: 'firms form partners for the dance but, when the music stops, they can change them. In these circumstances competition is still at work even if it has changed its mode of operation'. 36 See Lomi and Pattison (2006) for an empirical study of organization of production across multiple networks.

36

diversified connections. Others dimensions of 'ap-proximation' – i.e. cognitive,

organizational, social and institutional – among producers in a network may affect the

‘transfer of complementary pieces of knowledge’ (Boschma, 2005:71). On the other

hand, the presence of complementary pieces of knowledge and, thus, the possibility to

take advantage of them by networking implies the existence of differences and distances

in capabilities endowments. For two (or more) nodes of a producers’network ap-

proximate, their differences and distances must not exclude the possibility of

connecting/being compatible in at least one technological or productive capability

dimension. Thus, proximity is a concept which refers to the structural compatibility of

two (or more) different production organizations embedded in a network. Given the space

of capabilities C1 = [cij ] and C2 = [cij ] of two different production organization F1 and F2,

their proximity Φ is a function of the number of feasible connections among the fund

agents’ capabilities respectively owned by F1 and F2.

Φ : C1 = [cij ] C2 = [cij ]

The mapping between F1 and F2 can be extended to consider the entire set of nodes

embedded in a network form of production organization. By recurring to adjacency

matrices, we can analytically develop a matrix of proximity Φ. Here, each element

indicates according to a binary codification if fund of capabilities of one firm are

compatible with those of the others. For example, it will show if the machine owned by

F2 achieves the standard of precision or scale which are necessary for complementing in

production those machines owned by F1.

Below a certain level of proximity Φ < Φ, it is very unlikely that the two firms

will adopt a network form of production organization. An example would help here. Let

us consider a high-tech manufacturing firm F1 interested in splittering its production by

outsourcing the production of a set of components of the commodity A to another

manufacturing firm F2. To be sold in international markets the commodity A has to

satisfy certain properties and quality tests. In order to achieve these output standards, F1

37

has to be sure that the productive and technological capabilities through which it

performs part of the production process are compatible with those capabilities in use in

the firm F2. For example, as stressed above, F1 has to be sure that machines, workers etc.

owned by F2 are able to perform the production of certain components on time and

according to the required scale standards. Thus, outsourcing is not simply a problem of

cost reductions but mainly one of structural compatibility of capabilities in production.

This explains why, very often, producing in proximity requires investments in

transferring of tools, machines, mutual personnel training, drawings exchange, effective

knowledge transfer etc. At the end, all these strategies are meant to connecting and

making compatible different productive organization endowed with different capabilities

and to make possible the discovery and exploitation of complementarities among them.

For this reason, producing in proximity is a process which necessitates a continuous

investment in those skills which are needed for undergoing processes of structural

learning in network forms of production organization. To sum up, the concept of

producing in proximity refers to the fundamental dimension of structural compatibility

among producers in a network. By doing that, this concept highlights the need to consider

in a problematic way the bundle of relationships which realizes in multilayered networks

of producers – i.e. global production networks.

As for the second aspect – i.e. multilevel proximity, as the network form of

production comprises both horizontal relationships among production organizations at the

regional level as well as vertical relationship at the global level, different degrees of

proximity will be experienced by organizations according to their type of relationships.

As a matter of fact, networks of producers at the regional and global level overlap and

interact in a multilayered network structure. For example, complex and interlocking

clusters such as the industrial district in Emilia Romagna or the Baden-Wurttemberg

region (Piore and Sabel, 1984; Lorenzoni and Ornati, 1988; Best 1990; Pyke and

Sengenberger, 1992: Quadrio Curzio and Fortis, 2002; Becattini, 2004) are regional cases

of network forms of production organization. At the global level, the creation of

international networks of sub-contractors characterized by a certain degree of stability is

38

another example of close cooperation in a network form among increasingly specialized

organizations. As the network alignment approach has recently documented, both

regional and more global networks are actually strongly interconnected in a system

'where the relationships are many-to-many rather than one-to-the-next' (Dyker and von

Tunzelman, 2002:2). From a normative perspective, the development of technological

and organizational capabilities can neither derive only from ‘in-house’ technological

efforts by the firms nor simply from developing local ‘cluster knowledge systems’ which

remain closed to global production/technological networks. The development of

capabilities requires a pragmatic integration of ‘locally and nationally emerging networks

with global network structures’ (Kim and von Tunzelman 1998:1; Mc Gowan et al.,

2004: chapter 3; Iammarino et al., 2008). This integration will be responsible for the

emergence of new forms of production organization and will trigger new processes of

structural learning.

Concluding remarks

By integrating structural approaches to production with capability theories, the paper

outlines a capability theory of production. The black box of the production process has

been opened by describing it as a particular network of interrelated tasks through which

transformations of materials are performed according to different patterns of capabilities

coordination, subject to certain scale and time constraints. Capabilities are structurally

constrained, although an entire spectrum of virtual organizational and technological

arrangements is possible. These possibilities are generated by multiple relationships of

complementarity and similarity among structural components –i.e. materials in

transformation, productive agents and network of tasks – both at the firm and inter-firm

level. Discovering these possibilities is the very essence of a fully endogenous process of

learning – i.e. structural learning in historical time. Complementarities are essential

39

focusing devices in the continuous process of reconfiguration of the analytical map of

production relationships. In this process of change, problems of capability compatibility

and coordination arise both at the firm/industry and inter-firm/industry level. Different

forms of production organization have historically emerged and reappeared in response to

these coordination problems. Production organizations develop in historical time as a

result of exogenous circumstances as well as endogenous dynamics according to patterns

of specialization and diversification. In today’s global division of labour and knowledge,

the need for coordinating closely complementary but dissimilar activities has found a

response in the network-form of production organization. The effectiveness of these

networks in coordinating production and innovation activities is determined by the degree

of proximity among the agents embedded in the network and, thus, among their

capabilities.

The analysis of the internal structure of production and its change is essential for

investigating the structural dynamics of economic systems as well as for developing a

‘political economy of capability building’ (Cimoli, Dosi and Stiglitz, 2009: chapters 1

and 20). Structural economic dynamics identifies ‘those economic transformations that

explicitly account for the relative persistence of certain elements or relationships of

economic structure while other elements or relationships are subject to change’

(Landesmann and Scazzieri, 1996:3)37. At the micro-meso level – i.e. firms and local

networks, structural learning determines which components and processes in productive

structures will change or persist over time, both qualitatively and quantitatively. As a

matter of fact, structural transformations at the micro-meso level profoundly affect

structural dynamics of the overall economic system – i.e. macro level. However, there is

also a chain of causation which goes from the macro to the meso-micro level. The

possibility to influence and direct micro-meso structural dynamics through selective

policies is mainly in governments’ hands. According to Chang’s definition (1994:60)

industrial policies are policies ‘aimed at particular industries (and firms as their

components) to achieve the outcomes that are perceived by the state to be efficient for the

37 See also Baranzini and Scazzieri (1990)

40

economy as a whole’. Historically and across countries, selective industrial policies have

been main drivers of technological and organizational capabilities building (Chang, 2002:

2009; Andreoni, 2010b). These micro-meso dynamics account for the extraordinary

processes of structural change experienced at different points in history by today’s

industrialized countries.

The usefulness of a theory of production critically depends on its contribution to

the development of a new political economy of capabilities building. The capability

theory of production proposed above suggests two main research directions. First of all, it

highlights the advantages of re-embedding production theory into a structural economic

framework as well as to look at capabilities as structurally constrained. The need for a

‘re-structuralization’ of the industrial policy debate has been raised in recent debates

(Chang e Lin 2009) and some contributions (Chang, 2009; Lin 2010; Lin and Monga,

2010). Secondly, a capability theory of production stresses the importance of triggering

processes of structural learning through complementarity discovery both at the sectoral

and at the ‘super-sectoral’ levels. Industrial policies for the coordination of

complementary as well as competing investments, can activate processes of structural

reconfiguration of catching up economies and orientate their integration into global

production networks. Moreover, given the increasing pervasiveness of the network form

of production organization, industrial policies should focus on ‘connecting problems’

such as facilitating effective knowledge flows in producers’ networks; promoting sectoral

intersections and linkages; supporting adjustment costs; restoring industrial commons. In

sum, useful theories of production are necessary for designing and justifying new

policies. The final aim of these policies for capability building and accumulation is to

facilitate the discovery of new ‘worlds of possibilities’ and, thus, the emergence of ‘new

worlds of production’.

41

References

Abramovitz, M. (1986) ‘Catching Up, Forging Ahead, and Falling Behind’, Journal of Economic History, XLVI 385-406. Andreoni, A. (2010) ‘From Transition to Integration Through Institutional Innovation: The CENROPE as a model for cross-border capabilities development’, paper presented at the MyPhD Int.Conference ‘Where do we go from here?’, Bratislava 1-2 October 2010. Andreoni, A. (2010b) ‘More than a good night watchman – Rethinking the role of the state in the process of technological catching up’, Annali della Fondazione Einaudi, vol. 43, forthcoming. Andreoni, A. and Pelligra, V. (2009) Microfinanza – Dare Credito alle Relazioni, Bologna: Il Mulino. Antonelli, C. (1995) The Economics of Localized Technological Change and Industrial Dynamics, Boston: Kluwer Publishers. Atkinson, A. and Stiglitz, J. (1969) ‘A new view of technological change’, The Economic Journal, 79, 573–578. Babbage, C. (1832) On the Economy of Machinery and Manufacturing, London: Charles Knight. Baranzini, M. and Scazzieri, R. (1986), eds., Foundations of Economics: Structures of Inquiry and Economic Theory, Oxford and New York, Basil Blackwell. Baranzini, M. and Scazzieri, R. (1990) ‘Economic Structure: Analytical Perspectives’, in M. Baranzini e R. Scazzieri, eds., The Economic Theory of Structure and Change, Cambridge: Cambridge University Press. Becattini, G. (2004) Industrial Districts: A New Approach to Industrial Change, Cheltenham: Edward Elgar. Bell, M. (1982) ‘Technical Change in Infant Industries: a Review of the empirical Evidence’, mimeo, SPRU, University of Sussex. Bell, M. and Pavitt, K. (1995) ‘The Development of Technological Capabilities’, in I.U. Haque, Trade, Technology and International Competitiveness, EDI Development Studies, The World Bank, Washington D.C.: 69-101.

42

Bell, M. and Albu, M. (1999) ‘Knowledge systems and technological dynamism in industrial clusters in developing countries’, World Development, Vol. 27, No. 9, pp. 1715-1734. Best, M. (1990) The New Competition. Institutions of Industrial Restructuring, Cambridge: Polity Press. Best, M. (1999) ‘Regional Growth Dynamics: A Capabilities Perspective’, Contribution to Political Economy, 18, 105 – 119. Bellet, M., Colletis, G. and Lung, Y. (1993) ‘Introduction au numéro special Economie de proximités’, Revue d’Economie Regionale et Urbaine, 3, 357 – 361. Boschma, R. (2005) ‘Proximity and Innovation: A Critical Assessment’, Regional Studies, 39, 61 – 74.

Buenstorf, G. (2005) ‘Sequential production, modularity and technological change’, Structural Change and Economic Dynamics, 16, 221 – 241. Camagni, R. (1991) 'Local milieu, uncertainty and innovation networks: Towards a new dynamic theory of economic space', in, Camagni, R., ed., Innovation Networks: spatial perspectives, London: Belhaven Press. Cartwright, N. (1984) Nature’s Capacities and their Measurement, Oxford: Clarendon Press. Chandler, A. D. (1977) The Visible Hand, Cambridge, Mass. and London, England: The Belknap Press of Harvard University Press Chang, H-J. (1994) The political economy of industrial policy, Basingstoke: Macmillan. Chang, H-J. (2002) Kicking Away the Ladder, London: Anthem Press. Chang, H-J. (2009) ‘Industrial Policy: Can we go beyond an unproductive confrontation?’, a Plenary Paper for ABCDE (Annual World Bank Conference on Development Economics), Seoul, South Korea, 22-4, Revised: September 2009. Chang, H-J. and Lin, J. (2009) ‘Should industrial policy in developing countries conform to comparative advantage or defy it?, in Development Policy Review, vol. 27, no. 5. Cimoli, M., Dosi, G. and Stiglitz J., (2009) eds, Industrial Policy and Development. The Political Economy of Capabilities accumulation, Oxford: Oxford University Press.

43

Coase, R. H. (1937) ‘The nature of the firm’, Economica N. S., 4, 386 – 405. Dodgson, M. (1991) The management of technological learning, Berlin: De Gruyter. Dyker, D.A. and von Tunzelman (2002) ‘Network Alignment in Firms in Transition Countries: a survey of Hungarian and Slovenian Companies’, Working Papers ESRC One Europe or Several?, UCL and SPRU. Dosi, G., Nelson, R. R. and Winter, S. (2000) The nature and dynamics of organizational capabilities, Oxford: Oxford University Press. Foss, N. (2003) ‘Bounded rationality and tacit knowledge in the organizational capabilities approach: an assessment and a re-evaluation’, Industrial and Corporate Change, 12, 2, 185 – 201. Georgescu-Roegen, N. (1969) ‘Process in Farming versus Process in Manufacturing: A Problem of Balanced Development’, in Papi, U. and Nunn, C. (eds.) Economic Problems in Agriculture in Industrial Societies (Proceedings of a Conference of the International Economic Association, Rome, September, 1965) Georgescu-Roegen, N. (1970) ‘The economics of production’, American Economic Review, LX (2), 1 – 9. Georgescu-Roegen, N. (1971) The Entropy Law and the Economic Process, Cambridge MA: Harvard University Press. Georgescu-Roegen, N. (1976) Energy and Economic Myths, New York: Pergamon Press. Georgescu-Roegen, N. (1986) ‘ Man and Production’, in Baranzini, M. and Scazzieri, R. (eds.) Foundations of Economics, Oxford: Blackwell, 245 – 280. Georgescu-Roegen, N. (1990) ‘Production Process and Dynamic Economics’ in Baranzini, M. and Scazzieri, R. (eds.) The Economic Theory of Structure and Change, Cambridge: Cambridge University Press, 198 – 226. Hicks, J. (1969) A Theory of Economic History, Oxford: Clarendon Press. Hirschman, A. (1958) The Strategy of Economic Development, New Haven: Yale University Press. Kim, S-R. and von Tunzelman, N. (1998) ‘The dynamics and alignment of networks of networks': explaining Taiwan's successful IT specilization’, SEWPS No 17, Brighton: SPRU, University of Sussex.

44

Kogut, B. and Zander, U. (1992) ‘Knowledge of the firm, combinative capabilities, and the replication of technology’, Organizations Science, 3, 383 – 97. Iammarino, S., Padilla-Perez, R. and von Tunzelman, N. (2008) ‘Technological Capabilities and Global-Local Interactions: The Electronics Industry in Two Mexican Regions’, World Development, 36, 10, 1980 – 2003. Lall, S. (1982) ‘Technological Learning in the Third World: Some Implicationsof Technological Exports’, in Stewart, F. and James, J. (eds.) The Economics of New Technology in Developing Countries, London: Frances Printer, pp. 105-37. Lall, S. (1987) Learning to Industrialize: The Acquisition of Technological Capability by India, London: Macmillan Press. Lall, S. (1992) ‘Technological Capabilities and Industrialization’, World Development, Vol. 20, No. 2, pp. 165-186. 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. (1988) ‘Where do we go from here? Technical Change and Dynamics’, Cambridge Journal of Economics, 12, 169 – 177. Landesmann, M. and Scazzieri, R. (1996) Production and Economic Dynamics, Cambridge: Cambridge University Press. Langlois, R.N. (1992) ‘Transaction-cost Economics in Real Time’, Industrial and Corporate Change, 1, 1, 99 – 127. Langlois, R.N. (1998) ‘Capabilities and the Theory of the Firm’, in Foss, N. and Loasby, B. (eds.) Economic Organization, Capabilities and Coordination: Essays in Honour of G.B. Richardson. London: Routledge. Leontief, W. (1947) ‘Introduction to a Theory of the Internal Structure of Functional Relationships’, Econometrica, 15, 4, 361 – 373. Lin, J. (2010) ‘New Structural Economics. A framework for rethinking development’, Policy Research Working Paper, 5197, June. Lin, J. and Monga, C. (2010) ‘The Growth Report and New Structural Economics’, Policy Research Working Paper, 5336, June.

45

Loasby, B.J. (1999) Knowledge, institutions, and evolution in economics, London New York : Routledge. Lomi, A. and Pattison, P. (2006) ‘Manufacturing Relations: An Empirical Study of the Organization of Production Across Multiple Networks’, Organization Science, 17, 3, 313 – 332. Lorenzoni G. and Ornati O. (1988) ‘Constellation of Firms and New Ventures’, Journal of Business Venturing, 3, 41-57. Lundvall, B. A. (1992) National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning, London: Pinter. Marshall, A. (1920) Principles of Economics, 8th edition, London: Macmillan. Marx, K. ([1867]1972) Capital, a Critical Analysis of Capitalist Production, London, Lawrence and Wishart, vol I. Mc Gowan, F., Radosevic, S. and von Tunzelman, N. (2004) eds., The Emerging Industrial Structure of the Wider Europe, London and New York: Routledge. Mokyr, J. (2002) The Gifts of Athena: Historical Origins of the Knowledge Economy. Princeton: Princeton University Press. Morroni, M. (1992) Production Process and Technical Change, Cambridge: Cambridge University Press. Morroni, M. (2005) Knowledge, scale and transactions in the theory of the firm, Cambridge: Cambridge University Press.

Nelson, R.R. and Winter, S. (1982) An Evolutionary Theory of Economic Change, Cambridge, Mass.: Belknap Press

Pasinetti, L.L. (1959) ‘On concepts and measures of changes in productivity’, Review of Economics and Statistics, XLI, 3, 270 – 286. Pasinetti, L.L. (2007) Keynes and the Cambridge Keynesians, Cambridge: Cambridge University Press. Penrose, E. T. (1995) The Theory of the Growth of the Firm, 3rd edition. Oxford: Oxford University Press.

46

Piacentini, P. (1995) 'A Time-Explicit Theory of Production: Analytical and Operational Suggestions Following a “Fund-Flow” Approach', Structural Change and Economic Dynamics, 6, 4, 461-483. Piore, M. and C. Sabel, (1984) The Second Industrial Divide: Possibilities for Prosperity, New York: Macmillan. Pyke, F. and Sengenberger, W. (1992) Industrial Districts and Local Regeneration Geneva: International Institute for Labour Studies. Polanyi, M. (1958) Personal Knowledge. Towards a Post Critical Philosophy, London: Routledge. Poni, C. (2009) La seta in Italia, Bologna: Il Mulino. Quadrio Curzio, A. and Fortis, M. (2002) Complexity and Industrial Clusters: Dynamics and Models in Theory and Practice, Heidelberg and New York: Physica Verlag. Powell, W.W. (1990) ‘Neither Markets nor Hierarchy: Network Forms of Organization’, Research in Organizational Behaviour, 12, 1990, 295 – 336. Richardson, G. B. (1960) Information and Investment, Oxford: Oxford University Press. Richardson, G. B. (1972) 'The organisation of industry', Economic Journal, 82, 883–96. Robinson, J. (1953-54) ‘The production function and the theory of capital’, Review of Economic Studies, 21, 1, 81 – 106. Rosenberg, N. (1969) ‘The Direction of Technological Change: Inducement Mechanisms and Focusing Devices’, Economic Development and Cultural Change, 18, 1 – 24. Rosenberg, N. (1982) Inside the Black Box: Technology and Economics, Cambridge University Press. Rosenberg, N. (1976) Perspectives on Technology, Cambridge: Cambridge University Press. Rosenberg, N. (1979) ‘Technological Interdependence in the American Economy’, Technology and Culture, 20, 1, 25 – 50. Rosenberg, N., (1994) Exploring the black box: technology, economics, and history, Cambridge: Cambridge University Press.

47

Scazzieri, R. (1983) ‘The Production Process: General Characteristics and Taxonomy’, Rivista Internazionale di Scienze Economiche e Commericiali, XXX, 7, July, 597 - 611. Scazzieri, R. (1993) A Theory of Production. Tasks, Processes and Technical Practices,, Oxford: Clarendon Press. Scazzieri, R. (1999) ‘A Theory of resilient flow-fund linkages’, in Mayumi, K. and McGowdy, J. Bioeconomics and Sustainability, North. USA: Edward Elgar. Sen, A. K. (1960) Choice of techniques, Oxford: Basil Blackwell. Sen, A. K. (1985) Commodities and Capabilities, Amsterdam: North Holland. Simon, H. (1957) Administrative Behaviour. A Study of Decision-Making Process in Administrative Organisation, New York: Macmillan. Simon, H. (1962) ‘The Architecture of Complexity’, Proceedings of the American Philosophical Society, 106, 6, 467 – 82. Simon, H. (1969) The Science of the Artificial, Cambridge: MIT Press. Simon, H. (1991) ‘Bounded rationality and organizational learning’, Organization Science, 2/1, 125 – 134. Smith A. (1776) An Inquiry into the Nature and Causes of the Wealth of Nations, London: Strahan and Cadell Salais, R. and Storper, M. (1997) Worlds of Production: the Action Frameworks of the Economy, Cambridge, MA: Harvard University Press. Sabel, C. F. and Zeitlin, J. (1997) eds., World of Possibilities: Flexibility and Mass Production in Western Industrialization, New York: Cambridge University Press. Stewart, F. (1981) ‘International Technology Transfer: Issues and Policy Options’ in Streeten P. and Jolly R., eds, Recent Issues in World Development, Oxford: Pergamon Press, pp 67-110. Sylos Labini, P. (1995) ‘Why the interpretation of the Cobb-Douglas production function must be radically changed’, Structural Change and Economic Dynamics, 6, 485 – 504. Teece, D. (1980) ‘Economics of scope and the scope of the enterprise’, Journal of Economic Behavior and Organization, 1, 223-247.

48

Teece, D., Pisano, G. and Shuen, A. (1997) ‘Dynamic capabilities and strategic management’, Strategic Management Journal, 18, 7, 509–533. Von Tunzelmann, N. and Wang, Q. (2007) ‘Capabilities and Production Theory’, Structural Change and Economic Dynamics, 18, 192 – 211. Young, A. A. (1928) ‘Increasing returns and economic progress’, The Economic Journal, 38, 527–42.