Learning rates

We provide a new procedure for estimating the learning curves which are applied in integrated assessment model. As discussed in the publication (see below), the procedure is more robust to econometric biases than the standard method which was used previously.

Files required to use the module

All files required to use the module are available here.

Module elements

Input data

  • Data on electricity and energy prices, indicators for environmental policies and net capacities of wind turbines by country and by year: net capacities - all OECD from 1960.dta
  • Wind turbine instalation prices in US: INST. PRICE - US wind.dta
  • Mapping between country  names and country codes: Count country iso2 correspondance.dta

Description: The learning rates are computed using the historical data which are provided in the data sets above. The users can update the data sets (e.g. by including new observations from last years) or tailor the data set to a particular Integrated Assessment model by following the instructions in the STATA code.

Source code

  • STATA code: demonstration.do

Description: The STATA code processes the data and runs econometric model. The output is the estimate of the learning rate

How to use the module

Copy all the attached files into one folder and run the STATA code. If you wish to use your own dataset, adjusted to your integrated assessment model, follow the instructions described in the comments within the STATA code.


Witajewski Baltvilks, Jan, Elena Verdolini and Mssimo Tavoni 2015. "Bending the Learning Curve," Energy Economics, Elsevier, vol. 52(S1), pages S86-S99.