Innovation policy experimentation using an agent-based model of technological development by value networks in a multi-region industry
We engineer policies to stimulate technology development in a manufacturing industry using an agent-based model of value networks within a multi-region industry. Generally, industrial production is organized in value networks, where firms in upstream tiers supply components to firms in downstream tiers. We postulate that ties are increasingly intraregional with incremental innovation typical for the mature stage of the industry life-cycle, while ties are increasingly interregional to accommodate experimental recombination and discontinuous innovation in the decline and inception stages of the industry life-cycle. Policy instruments should create conditions favorable for regional performance in either of the two distinct industry life-cycle phases. We present an agent-based model based on Pyka, Gilbert & Ahrweiler (2007) in which agents decentrally form multi-tier production networks that produce products for the end-consumer market. Production knowledge is modeled as a capability taxonomy with an expandable set of abilities per capability class. These abilities are the production recipes to transform input products acquired at upstream tiers into output products sold at downstream tiers. Agents continuously experiment with their abilities to improve the product they produce, but the increase in performance of products gradually tapers off. In case agents repeatedly discover inferior products, they look for new combinations of capabilities, both within and outside the set of their current value network partners. This radical capability experimentation may yield new capabilities, thereby new possible abilities, hence possibly new (end-)products that can be produced. Firms either collaborate with other firms, collaborate with a research institute, or vertically integrate to access capabilities that may or may not 'unlock' new capabilities. We introduce a spatial limit to both the production network partnership (and thereby ability hill-climbing) and the innovation network partnership (and thereby capability discovery search). To understand which technology search heuristics are to be stimulated and what research infrastructure is to be put in place, we study technological progress under different distributions of capabilities (over regions, over public pools, and over firms in the value networks) and different (spatial characteristics of) search heuristics used by the agents. Finally, we initialize our agent-based model to reflect the particular cases of the biotechnology industry in Vienna and automotive industry in the Stuttgart region. Using the empirical distribution of capabilities over agents and regions, we study technological progress in both industries for the different scenarios for the (spatial characteristics of) search heuristics and for different research infrastructures for capability pooling.
When requesting a correction, please mention this item's handle: RePEc:wiw:wiwrsa:ersa13p225. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Gunther Maier)
If references are entirely missing, you can add them using this form.