1In their outstanding book Antonelli et Barbiellini Amideo (ABA thereafter) offer two perspectives which have different aims (although complementary): 1) providing materials to explain the rate and the direction of economic growth in Italy, 2) in the context of Italian economy building a theoretical frame which enables to assess how Technological Change can fuel economic growth, or how Knowledge creation affects Wealth production. The two parts of the book are different as far as methodology is concerned. The first part (“The localized generation and exploitation of technological knowledge and innovation”)draws hugely on the theoretical analysis of localized technological knowledge patiently built up by Cristiano Antonelli in the past years. The roots of his approach have to be found in the Marshallian approach of the firm (see for instance chapter 6) and the Schumpeterian intuition related to the necessary creative response of the agents to the signals coming from their environment. With the seminal contributions from Arrow and David to the Economic of knowledge these roots give the theoretical micro foundations of the book (better labeled in their updated form as a “still to build” theory of economic complexity). The second part (“The Italian evidence”) provides a lot of statistical evidence and econometrical calculations on the productivity, RandD expenditures and the innovation drivers mainly at the sector level (and Italian regions level as well). Put in the Growth accounting model they enable to discover the main factors pushing or pulling the growth of the total factor productivity and to test some authors’ideas. So the aim of the book in last instance is to provide a theoretical and empirical framework for understanding the postwar Italian growth.
2The book is the outcome of a long term research project led by the Research Department of the Bank of Italy. It opens on a central issue: the intriguing puzzle of the Italian growth from the aftermath of the Second World War to 1992 (a particular date in the process of the European economic integration characterised by the so-called “Single Market Program”). For this time period, a lot of evidence has clearly shown a fast increase of total factor productivity (TPF) that is the conventional definition of economic growth, pushed by technological and organizational innovations. Surprisingly, the Italian economy is characterized by low levels of R&D expenditures in comparison with the countries having the same level of economic development. As a consequence, there is a puzzle to be solved: why does such a strong rate of total productivity growth coexist with “low” levels of R&D expenditures?
3This kind of puzzle could and should be explained by factors related to the different competitive or complementary drivers of productivity growth, among the more relevant which can be quoted:
The implementation of new methods of work organization. In the first part of the time period under observation (1950-1992), the diffusion in Europe of taylorist/fordist methods in particular in the consumer goods industries has significantly productivity (Boyer and Petit, 1981). This implementation can be achieved without R&D investments. Such a phenomenon deeply interacts with firm size and enables firms to achieve or obtain economies of scale (a growth determinant at the core of the Kaldor-Young tradition of productivity growth). ABA do not retain this factor (we will see later why) that can works with other drivers including:
Structural change, which has already been considered by Denison (1967) in his very first study on the sources of US growth, plays a certain role in this story. ABA dedicate the entire chapter 10 to this factor which is considered as having a “crucial influence” (page 124). It is correlated to the emergence of a strong domestic machinery industry in the mid-1960s (page 127 and on).
The transfer of technology through the import of capital goods or licensing with foreign players. With respect to this factor, enabling the diffusion of new technologies born abroad and the modernization of the domestic stock of capital goods, ABA note that it was during the decades that followedthe Second World War. However, with the emergence of the Italian machinery sector, it ceased to be important. Lastly, the country has reversed the trend and can now export its own technologies (all these aspects are discussed in chapter 10). In accordance with this view the authors note that “technological payments should be considered an integral and crucial part of the national innovative efforts; a complementary factor to R&D, an important input of Italian industry localized innovation processes” (page 111).
4These factors matter and have real impacts (particularly once articulated each other) on the global productivity growth, but do not really provide a powerful explanation. In fact, the “true solution” to the puzzle is given in the introduction, and discussed at length in the chapter 8 (“Technological change in a distributed innovation system”) and of course sets up the framework of the book. The solution to the puzzle is “found in the systemic congruence that has emerged in the Italian economy between the organization of the generation of technological knowledge and the traditional characteristics of the economic and industrial structure” (page 1). The important message is given further on: “The innovative ability of Italian firms was based upon the systematic valorization of user–producer interactions between upstream producers of capital goods and downstream producers of consumer goods, within Italian districts, and the systematic development of localized learning rather than on the traditional ‘corporate model’ based upon mechanisms of formal research funded and performed internally by large firms” (page 2). If we want to be a little polemical we could jokingly say “the Italian district is the solution”. It is a part of the story, but too short. The Italian district is a variety of marshallian industrial districts, ABA show that it is more complex in terms of industrial structure. In effect ABA paint a pictureof the district that does not seem to be the ‘dominant design’. They focus their study on the ways with which the players in the districts produce, learn, and exchange knowledge. They develop a particular empirical model of localized technological change cum pecuniary knowledge externality. Its theoretical foundations are laid in chapter 7 (possibly being the core of the book as far as theory is concerned): “Pecuniary knowledge externalities are found if, when and whereknowledge- utilization costs differ sharply from its generation, andwhen knowledge-governance at the system level is effective and theknowledge-governance mechanisms are highly efficient” (page 54). Using a Knowledge production function specific to the firm coming from the works by Nelson (1982) and Weitzman (1996, 1998) in which the two main complementary inputs are the internal and external Knowledge. When pecuniary knowledge externalities apply the equilibrium technique will consist in using external knowledge more than internal knowledge. As a consequence, firms will produce more technological knowledge than in a system where external knowledge has higher costs. Adding to this model developed in the context of equilibrium microeconomic textbook firm a Grilichesian production function in which knowledge enters as an argument, it has been shown that total factor productivity levels and their growth can depend upon the divergence between the equilibrium costs of external knowledge and the actual ones. As a consequence, ABA obtain the basic proposition that total factor productivity levels increase with pecuniary knowledge externalities. As noted by the authors, “As long as there are pecuniary knowledge externalities, the typical system dynamics, stemming from the positive feedback generated by knowledge non exhaustibility and knowledge complementarity, implemented by good knowledge governance mechanisms, are at work at the system level” (page 58). The scene is set.
5The Marshallian Economics tradition has identified two types of externalities: technological externalities and pecuniary externalities (see the seminal definitions by Scitovsky, 1954). Technological externalities consist of direct interdependence among producers. The location of the hives of X close the fields of Y done that the bees can forage in his fields sets up a well known example of technological interdependence affecting the (honey) production function of X. As seen through this example pecuniary externalities consist of indirect interdependence not mediated by the market mechanisms (Y does not sell his pollen –the unpaid factor - to X). With pecuniary externalities instead, interdependence between agents takes place via effects on price mechanisms. Prices take explicitly into account the external forces affecting production functions. Pecuniary externalities are positive when the market price of production factors happens to be lower than equilibrium levels due to the effects of market interactions among agents, and conversely for negative pecuniary externalities. The foundations of the book argue pecuniary knowledge externalities play a key role in shaping the direction and the rate of technological change, and the dynamics of the convergent processes of knowledge generation and exploitation that lead to the emergence of systems of innovation. It is this type of dynamics that structures the mechanisms of the so called “Kaldor-Verdoorn laws”. A firm in a country having a higher growth rate on a medium-long period will benefit from stronger and more specialized capital-goods sectors.
6Some remarks about this model of localized technological change cum pecuniary knowledge externality: First, ABA must be congratulated for exploring this difficult path which aims to find a theory, a technological change, with the tools of the microeconomic approach to the firm. Their brilliant analysis using some assumptions of the theory of complexity raises numerous questions. In the frame of this note only some of them among the more crucial questions can be discussed. For instance why is there a double-step? Why does the firm firstly choose the proportion of external Knowledge in the Nelson-Weitzman Knowledge production function and, then through the traditional production function its quantity of other inputs? Why could not the firm build up the choice of its factors of production (tangible capital, labour, internal Knowledge, external Knowledge – that is not free as underlined in the book) at the same time? Another point would deserve more attention. It is a flow that is the output of the Nelson-Weitzman Knowledge production function. By contrast, the conventional schemes of economic growth theories retains in general a production function with a “capital of Knowledge”, an accumulation of Knowledge over several time periods, a stock of Knowledge but not a flow (Griliches, 1995) partly inherited from the past. However it is true that Griliches (1995) considers equally an alternative formulation in which the rate of growth of productivity is explained by the intensityof the net investment in R&D. What ABA did to some extent. It seems to us that ABA are exploring a very fruitful analytical field when they want to link in a coherent manner a Nelson-Weitzman Knowledge production function and a production function with a “capital of Knowledge”.
7On the empirics of pecuniary knowledge externalities let us return to the industrial districts where the economic play is of a crucial importance. It is worth noting that the amount of pecuniary knowledge externalities made available by the localized pools of technologycan stimulate innovation in relation with the demand pull incentives. Very important (if not essential) in the argument put forth by ABA: the vertical user–producer relations between upstream producers and downstream users of capital goods pull the firms to increase the capital intensity of their production process and to implement more process and product innovations. In the space of the Italian districts, there is a geographical and relational proximity between users (downstream firms that produce consumer goods) and producers (upstream firms that produce capital goods) that favors Knowledge creation and circulation. At this point the authors point out; “The supply of new idiosyncratic inputs by upstream producers pushes downstream users to direct their technological change towards an intensive use of them, also in order to increase knowledge appropriability. The increased derived demand for these idiosyncratic products activates demand pull effects upon the rate and direction of technological change in upstream industries” (page 77). As a consequence,there is a very positive interplay between the dynamics of division of labor, the creation of new intermediary markets, the rate of technological change and the introduction of innovations: “The creation of new intermediary markets and the increased specialization of downstream firms in well- defined production layers provided opportunities for the bottom- up introduction of new dedicated technologies, as much as the introduction of technological innovations in the upstream chemical and mechanical industries provided an opportunity to identify new downstream market niches” (page 80). In their analysis the supply of pecuniary knowledge externalities by upstream innovating firms providing opportunities for the creation of new products in downstream applications play a key role. In turn, innovative users push upstream producers to introduce innovations. In sum, there is a dynamic close to that considered by Nicolas Kaldor but at a regional level. Another aspect deserves attention here. This biased technological change dynamic (due to the implementation of capital intensive process innovations) is the result of knowledge exploitation strategies stemming from the crucial role of user–producer relations in the generation of technological knowledge. Interestingly the specific direction of technological change is determined by the mode of innovation generation. It is a very fascinating new approach which links direction, rhythm, and a mode of generation of technological progress.
8It seems to us that a bridging might be undertaken with two institutionalist perspective on growth and technical progress; the French regulation school and the insights given by regional innovation system approach.
9French Regulation theory discusses economic change through two central concepts: regimes of accumulation and modes of regulation. The accumulation regime corresponds to a pattern of (more and less) regular growth. The mode of regulation represents the procedures and social behaviour supporting the accumulation regime and ‘ensure the compatibility over time of a set of decentralized decisions, without the economic actors themselves having to internalize the adjustment principles governing the overall system’. (Boyer, 1990; Petit, 1998). Five basic types of institutions set up the core of the mode of regulation: the forms of monetary constraint, the configurations of the wage-labour nexus, the forms of competition, the modalities of insertion into the international regime,the forms of the State. In such an approach a central question is to find the regime that will succeed the famous fordist system sustaining the high economic growth in the aftermath of the Second World War. While ABA do not develop their study in the frame of the regimes of accumulation, their analysis clearly suggests there exist a specific means for producing self reinforcing growth. Is this new regime specific to the Italian productive system? Does it have some likeness with the Scandinavian clusters exemplified by Porter? Is it a type of a neoschumpeterian regime of growth delineated by Boyer (2002)? From this point of view ABA book raises issues to the French Regulation school It seems to us that another (and may be better) bringing together should be done with the approach by Piore and Sabel. (1984) onto the positive impact of small and medium sized firms agglomerated in the same geographical location.
10The topics dealt with in the book deserve to be analyzed with the lens of regional innovation system (Asheim and Isaksen, 2002) or local productive system Courlet (2002). The space is missing in this note to explore the proximity with these approaches.
11For ABA, this type of dynamic along chains of specialized firms working in the Italian districts is the core explanation of the puzzle which account for the Italian productivity growth. In effect with the districts they find a strong capacity of innovation but a not so high R&D intensity. At this stage, it is worth noting that there are two phenomena that tend to underscore the level of R&D investments at the district level: 1) due to the fact that only people who workfull time in R&D activities are recorded as human R&D resources (a consequence of the standards from the famous Frascati Manuel), it may be the presence of a large number of small sized firms – that is a characteristic of the industrial structure of the district - tend to underestimate the volume of R&D expenditures recorded by the official statistics (acknowledged as well by Griliches, 1995), 2) ABA acknowledge in the districts the importance of the “soft innovations based upon design and fashion” (page 79). Design and related activities are secondary drivers of product innovation in downstream firms enabling to make a difference on the domestic and foreign competitive firms, but they are resource-consuming. For the upstream firms, design is also a crucial function (likely more than R&D) that drives innovation in new capital goods (machinery). We know that the expenditures related to design activity (and more generally creativity) are not recorded as R&D expenditures through the accounting standard of Frascati Manual. We suppose these expenditures are high in the Italian districts. In my opinion, if we correct this bias in the R&D amount due to the size of firms and retain another indicator as a relevant measure of technological investments (for instance the innovation costs including Design expenditures), it might be that the total amount of innovation costs in Italian industry would be higher and the “puzzle” less evident.
12In the type of analysis developed by ABA there is no place for large firms. For instance the “national champions” Fiat and Olivetti are almost never quoted (only one time for the first and two times for second in the footnotes). In the time period under analysis, the large multinational firms that innovate persistently emerge and prosper partly because of the creation and the enlargement of the European common market. It may be the impact of their growth on innovation and productivity trends which deserve much more attention (in relation with the implementation of taylorist/fordist methods as seen previously). ABA give more importance to the marshallian approach of industrial agglomeration (emended a little by the porterian view on industrial innovative cluster) than the chandlerian model, which, by contrast, stresses the crucial role of (very) large firms as a vector of innovation (see for instance Pavitt and Patel, 1995).
13Interesting (and uncommon) econometric calculations (chapter 12) tend to confirm the fruitful thesis articulated around two proposals: 1) the generation of innovations by downstream users is supported by the flows of pecuniary externalities from upstream, 2) A demand pull mechanism exists which is fuelled by the growth of the downstream sectors that generate a persistent upstream flow of technological innovations (it should be noted that such circular mechanisms are taken into account by the studies which are investigating the field of persistent innovation). As far as the important estimations set out in chapter 12 are concerned, one remark must be made. Firstly, if design and creativity are very important, particularly for the downstream sectors, it would be relevant to add in the relations to be estimated a variable “capital of design and creativity” built up with data available on the designs and models (at the Italian IP office or at the European office). This variable should explain the total factor of productivity directly for the downstream users and indirectly for the upstream producers (through the mechanisms pictured by ABA).
14In the end, this is an excellent book that is absolutely necessary for academics wishing to better understanding the complex relationship between TFP and technological change underlying the Italian growth. This is an essential book essential for the community of researchers who are trying to track how and when the forms of industrial organization (or the economic institutions) contribute to the creation and diffusion of “useful for growth” Knowledge. A good book for those who think that behind the “growth accounting models” there are the true roots of the engine of growth.