Learning by doing vs. learning by searching in models for climate change policy analysis

Most of the predictions and conclusions in the climate change literature have been made and drawn on the basis of theoretical analyses and quantitative models that assume exogenous technological change. How do these predictions and conclusions change if we endogenize technical progress? In this paper we consider two different drivers of technological change – Research and Development (R&D) and Learning-by-Doing (LbD) – and we embed them into the popular Nordhaus and Yang’s RICE model. We then use the corresponding two model versions to simulate different policy scenarios and compare the results focusing on consumption, physical capital, emissions abatement rates, and R&D expenditures. Our findings suggest that R&D-driven and LbD-driven technologies lead to quite similar dynamic patterns of the relevant variables we analyze. However, the greater flexibility enjoyed by agents who are able to optimally choose R&D expenditures seems to imply more welfare relative to the LbD case.