Dynamics of Technological Development in the Energy Sector

This paper reviews the literature on trends of technological improvement, focusing on the energy sector. We discuss the extent to which past trends can be used to predict the future improvement paths of technologies. The historical trends for certain technologies, such as wind and photovoltaics, have been much more regular than those of other technologies, such as nuclear fission or natural gas. Reasons for this include different degrees of dependency on scarce resources (which is high in the case of natural gas), as well as technology improvement drivers other than cost (such as a push to increase safety in the case of nuclear fission). Data from the United States show that retail electricity prices have fluctuated over the last forty years, but with no clear increasing or decreasing trend. In contrast the cost of several renewable technologies has dropped considerably; for instance, the cost of photovoltaics has dropped by more than two orders of magnitude during that same period. A blind extrapolation of historical trends suggests that the cost to achieve parity is not prohibitive, though we stress that there are large uncertainties involved. In an effort to better understand the reasons for these trends, we review theories for the functional form of technological improvement curves and discuss how this problem can be understood in terms of portfolio theory.