Learning curve with input price for tracking technical change in the energy transition process

Abstract

Abstract Learning curve has been recognized as an effective tool for tracking the technological change and widely used in assisting policy design for energy technologies. But uncertainties underlying input price and learning rate would greatly influence the accuracy and reliability of learning curves. Regression to a learning curve using lead-acid price and production (between 1989 and 2013) yields a poor fit performance. This study then integrates cost decomposition with learning curve to remove the influences from input price changes in lead-acid batteries. Total lead-acid batteries' cost is decomposed into materials cost and residual cost. Materials cost is calculated through multiplying material consumption and its price to isolate the influences from input price changes. Learning curve is then constructed to trace the declines of materials consumption and residual cost. Regression with the improved learning curves performs better than a standard one, with the fit performance improved greatly. The study also indicates a learning effect in the material consumption in the long term. Integrating cost decomposition into learning curve is then proved to be a meaningful trial for mitigating the bias from input price changes. Learning rates of mature technologies such like lead-acid batteries indicate the importance to understand the cost reduction in them, which might significantly affect the policies’ performances for emerging ones.