After more than one year collaborating with the University of Otago on machine learning models to better predict corporate carbon emissions, we’ve published a working paper in the United States Association for Energy Economics. You can read it here.

Summary:

• Big differences in various data provider (Bloomberg, Refinitiv/Eikon, and ISS) estimates of Scope 3 depending on the methodology.

• The application of machine learning algorithms can improve the prediction accuracy of the aggregated Scope 3 emissions (up to 6%) and its components, especially when each category is estimated individually and aggregated into the total Scope 3 emissions values (up to 25%).

• Overall accuracy of predictions is low (~70-80%) however it is easier to predict upstream emissions than downstream emissions. Prediction performance is primarily limited by low observations in particular categories, and predictor importance varies by category.