Completing the SME carbon profile: scalable prediction of scope 3 emissions

Despite growing recognition of significance, a business’s Scope 3 emissions remain rarely measured and as a result are poorly understood. This situation is particularly common amongst small and medium-sized enterprises (SMEs), which face additional obstacles to emission measurement. With this paper, we present a transaction-based approach to facilitate SME Scope 3 engagement. Using financial transaction data for 150,000 + UK SMEs, we produce spend-based Scope 3 estimates across key Greenhouse Gas Protocol categories. We then fit a series of hierarchical regression models to both quantify and identify firm-level Scope 3 emissions, with minimal user inputs. We find that this approach is effective in predicting the upstream emissions of both purchased goods & services (RSQ = 0.87) and fuel and energy-related activities (RSQ = 0.89 and 0.72), whilst weaker for more targeted categories such as business travel. We also find a small number of recurrent industry hotspots tend to account for 75% of a firm’s upstream emissions. By leveraging objective, standardised data to estimate emissions, this method provides a low-input alternative to costly micro-studies for generating Scope 3 insights, extending the visibility of emissions beyond a firm’s direct operations, revealing emission hotspots and supporting the development of value chain decarbonisation strategies.