FIGURE 5. Technological uncertainties: learning rates (push) and market growth (pull). SOURCE: Nebojsa Nakicenovic, International Institute for Applied Systems Analysis, presentation given at the Workshop on Assessing Economic Impacts of Greenhouse Gas Mitigation, National Academies, Washington, D.C., October 2-3, 2008.

FIGURE 5. Technological uncertainties: learning rates (push) and market growth (pull). SOURCE: Nebojsa Nakicenovic, International Institute for Applied Systems Analysis, presentation given at the Workshop on Assessing Economic Impacts of Greenhouse Gas Mitigation, National Academies, Washington, D.C., October 2-3, 2008.

current technologies have instead seen limited improvement despite continuous investment. Therefore, Nakicenovic and Nordhaus both recommended doing sensitivity analyses with the learning off, to help bound expectations.

Finally, Nordhaus described a third approach, the Romer model (Romer, 1990), which in his view is the right kind of model. It has conceptual and data problems that need serious work, but it has an explicit link between R&D and other inputs and the technology outputs. David Montgomery agreed that a Romer-type model may be the appropriate one to use, but offered three concerns. First is that the process of basic research is not very clear or predictable. Second, we do not entirely know how efficient the market for innovations is. Third, it is not easy with this model to determine where the levers might be to influence the rate or direction of technological progress, in order to reduce GHG emissions, particularly given that the influence of prices on R&D decisions is not well understood.

Nakicenovic added that the lack of data may be the major constraint to using a Romer-type model. He pointed to a recent report that Germany and some other countries in Europe are expecting to increase energy R&D efforts after more than a decade of drastic declines—trying to understand what drives this apparent inducement of technological change will be key to modeling issues. Richard Newell seconded the notion that empirical data is a critical limitation and suggested that there is a major need for more work in theoretical development of ways to model technological change. He and colleagues are currently working on some aspects of this, such as understanding market imperfections and the effects of spillovers.

Skip Laitner concurred that the Romer model was an appropriate beginning point but cautioned that modelers still tend to have an outdated view of technology and ought to improve their understanding of 21st century technologies. Inja Paik noted that the OECD has done a considerable amount of work in examining national innovation systems, how countries organize R&D resources to generate knowledge, and how diffusion of knowledge eventually contributes to their GDP. This work seems to have implications for the Romer model approach. Newell added that it continues to be difficult to evaluate prospective benefits from R&D investments without being able to model these in much more detail than is currently possible.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement