Modeling technological change under increasing returns and uncertainty

The aim of this paper is to analyze methodological challenges involved in modeling of endogenous technological changes with increasing returns and uncertainties by using stylized versions of models. Realistic versions of these models are analytically intractable making it difficult to comprehend the interplay of different assumptions on their outcomes. We demonstrate path-dependences of myopic evolutionary approaches, the infeasibility of straightforward "trial-and-error" processes, and the need for adequate long-term policy assistance. We also show why increasing returns and uncertainties radically offset the rationale for postponed investments in new technologies and how stochastic models cope with systemic risks implicitly induced by interdependencies among uncertainties, technologies, the structure of models, and decisions. The paper demonstrates possible misleading character of alternative models of uncertainties. It shows the need for proper modeling of long-term random horizons, corresponding discounting, security constraints and requirements of robustness by using systemic valuations and "distribution free" stochastic programming/optimization.

VIEW CONTENT