Experience Curves for Electrolysis Technologies

Abstract

Given the rapid increase in green hydrogen research funding and the hopes that this will help drive cost reductions, it is important to incorporate the latest RD&D spending into the estimation of the learning rate for electrolysis technologies. Thus, we develop a two-factor experience curve model with spillovers and economies of scale that allows us to estimate learning rates for both alkaline and PEM electrolysis technologies using both global- and country-level data from OECD countries. Our research strategy allows us to mitigate estimation or omitted variable bias from ignoring technology-push measures, unobserved country-specific characteristics, and knowledge spillovers. Using an OECD cross-country dataset over 2000-2022, we estimate global learning-by-doing rates of 17.5 %-46.8% and global learning-by-researching rate of 9%-42.3% for electrolysis technologies after incorporating learning parameter estimates into the progress equation. When we allow for spillovers, we find a linear relationship between PEM technology and alkaline technology improvements. Based on our OECD panel dataset, which incorporate more observations, we estimate learning-by-doing rates of 0.6%-9.4% and learning-by-researching rates of 4.0%-19.9%. In addition, country-level electrolysis cost is reduced by about 28% for the sample period 2000-2022 because of global experience spillover effects. Therefore, our empirical findings suggest that the role of technology-push measures remains critical for promoting and achieving cost improvements of electrolysis technologies. Furthermore, the absorptive capacity of a country should be improved to maximise the benefits of spillovers from global learning.