Introducing errors in progress ratios determined from experience curves

Abstract

Progress ratios (PRs) derived from historical data in experience curves are used for forecasting development of many technologies as a means to model endogenous technical change in for instance climate–economy models. These forecasts are highly sensitive to uncertainties in the progress ratio. As a progress ratio is determined from fitting data, a coefficient of determination R2 is frequently used to show the quality of the fit and accuracy of PR. Although this is instructive, we recommend using the error σPR in PR, which can be directly determined from fitting the data. In this paper we illustrate this approach for three renewable energy technologies, i.e., wind energy, bio-ethanol, and photovoltaics.