Make, Buy, or Re-design - The impact of digital technologies on the organization of production

The technologies which characterize the 4th industrial revolution may offer unprecedented possibilities of fast and radical re-design of products in such a way as to deeply modify the number of parts and components and the interdependencies among them. For instance, General Electric used 3D printing technology to re-design core components of its jet engine: the catalyst went down from 855 components to 12, the fuel nozzle from 20 components to 1. In general, although they are not yet so widely used, 3D printers allow for the production of objects with very complex shapes and multiple materials which before could only be obtained as assembly of multiple pieces.
Besides the obvious technical improvements that this technology may offer in terms of higher
robustness, lower weight, simpler maintenance, etc., additive manufacturing may also have important organizational implications. Traditionally we think that a firm in need of component faces the choice between producing it in-house (the make option) or going to the market and buying it from an independent supplier (the buy option). For instance, transactions costs theory puts this choice of making vs. buying at the core of organizational theory and proposes an explanation of why and when the making solution is superior to the default option of buying (Coase, 1937, Williamson, 1975). When additive manufacturing can be employed, a third option seems to appear: re-designing the product in such a way that the component is no longer needed or that the interdependencies among components are radically modified. Similar examples can be found also for other technologies of the 4th industrial revolution, such as robotics, internet of things, blockchain technologies and platforms.
Indeed, the option of redesigning existed also in the past: the product network of such complex product systems as, for instance, automobiles underwent substantial modifications, but, typically, re-designing was only a long-run option, allowed by process innovation, requiring considerable investment in physical capital and substantial modification of the production line. New technologies such as additive manufacturing make re-designing, at least in some cases, a viable short-run strategic option.
What are the implications for organization theory? Standard theories, as exemplified by transactions costs theory, typically consider the product network as exogenously given. Focusing only on the costs of transacting, TCE excludes from the analysis the possible interactions between organization and technology. Notably, if the product network can be endogenously modified, industry architecture is no longer the outcome of transaction costs minimization for a given product network (Langlois 1992, Pagano 1992, Jacobides & Winter 2005). Moreover, TCE only focuses on whether a specific edge of the product network should be managed by either a market or a hierarchy (or some intermediate hybrid). Indeed, TCE does not consider the network in its complexity, but only one of its edges in isolation, implicitly assuming that the optimal governance of the network will be given by the simple combination of the optimal governances of each of its edges, assuming away the possibility of complex
interactions among different edges and the deriving instability of organizational arrangements
(Marengo, 2020).
In this paper we propose a computational model, derived from the growing literature on NK fitness landscape, which studies the organizational implications of a plastic product network. Our model analyzes the interactions among three networks: 1) a product system which consists of the network of interactions among the components of a product; 2) an organizational network which consists of organizations producing subsets of components (according their degree of vertical integration) and buying or selling the remaining ones on markets; 3) a value network which governs the distribution of the value produced and where separate organizations may form coalitions in order to increase joint surplus.
In our model we analyze the interactions among all the three networks and, in particular, how changes of the product system network induce changes in the organizational and value networks and how firms can strategically modify the former in order to acquire better position in the second and more value in the third.
The product network represents the technology, its vertices are components of the product systems and a directed edge from components A to component B indicates that the production of the latter need the former (Arthur 2009, Brusoni, Prencipe, Pavitt 2001, Baldwin 2008,). The product network is modified whenever the number of components and/or their connections are changed, for instance thanks to the introduction of such technologies as 3D printing or blockchain.
The organizational network derives from some aggregation of the product network: vertices are
organizations which produce some subset of the product component and directed edges are transactions taking place between different organizations. The organizational network is modified whenever existing organizations either integrate or disintegrate vertically or whenever organizations exit or enter the industry.
The value network represents how the value generated by a product system gets distribute among the organizations forming the corresponding organizational network. Value capture theory (see the excellent survey in Gans and Ryall, 2017), analyzes how the value created in a product system is distributed among the participant firms. This stream of research typically determines lower and upper bounds to the amount of value individual firms or coalitions thereof can appropriate by comparing different configurations of the network, i.e. networks in which the firm or coalition under consideration is either present or absent or connects to a different subset of firms. In our model, the value network is modified whenever organizations form new coalitions or modify the existing ones.

In this paper we provide a detailed study of interaction among these three networks and, in particular, we analyze how changes in the product network influence the structure and stability of the other two networks.