Incorporating Forward-Looking Data in Probabilistic Analysis of Net-Zero Commitments

In this paper, we leverage state-of-the-art natural language processing approaches to assess the articulation and implementation of corporate net-zero transition plans. We use the retrieved transition-relevant data to enhance a probabilistic analysis of corporate net-zero commitments. In a Bayesian net, we combine the probability of staying below the net-zero budget with an assessment of the net-zero transition plan to obtain a forward-looking adjusted probability that accounts for both past emissions reduction efforts and future transition plans. We find that in the multisectoral dataset of consideration, the adjusted probabilities are lower than the original ones for most of the companies, which might indicate inferior climate performance as in the assumed climate scenario or also a potential switch to a worse climate scenario.