Posters

This study investigates the influence of extreme weather events, such as floods and wildfires, on pro-environmental beliefs and behaviors in Australia, with a focus on actuarial implications. It examines how these events affect climate risk mitigation, adaptation strategies, and sustainable investments. By integrating decision-making, psychology, climate science, and actuarial studies, we aim to disentangle the effects of climate change on various decision-making processes. Using longitudinal datasets, this research assesses whether extreme weather events impact preference formation and translate into climate-conscious choices. This project promotes interdisciplinary collaboration and identifies key datasets, enhancing actuarial models and strategies for managing climate risk and promoting sustainable investments. We utilize data on voting preferences, political alignments, insurance claims, and climate risk mitigation strategies to evaluate the long-term impact of climate risk. The results provide valuable insights for actuaries, policymakers, and community leaders to foster resilient and climate-conscious communities.

In this software paper, we outline the development and potential of Generative Pre-trained Transformer (GPT) models for integrating Environmental, Social, and Governance (ESG) factors into the investment management process, particularly for Thai stocks. As sustainable and responsible investing gains momentum, this study explores how GPT models, known for their advanced natural language processing capabilities, can revolutionize ESG integration. We examine GPT's ability to analyze and interpret vast amounts of ESG-related data, aiding in more informed and data-driven investment decisions. Our software covers applications of GPT in ESG investment strategies, such as materiality assessment, data analysis, integration strategy, and risk management. The paper also delves into the challenges and opportunities presented by GPT in sustainable investing, emphasizing its impact on enhancing investor engagement, ensuring compliance, and contributing to the broader objectives of sustainable financial markets.

This study investigates the incorporation of Environmental, Social, and Governance (ESG) factors into Thai stock market analysis, emphasizing the Thailand Sustainability Investment (THSI) platform. Central to this is the integration of a'best minus worst'derived ESG factor into the Fama-French three-factor model, serving as a proxy for ESG impact on stock returns. The research employs a homogeneity covariance matrix test to establish the ESG Best Minus Worst (BMW) factor's distinctiveness from traditional market factors. Results show the ESG BMW factor's unique covariance structure, enhancing existing models and highlighting variability due to different rating agency methodologies, thus impacting investment decisions. The study finds the modified Fama-French model with the ESG BMW factor more effective than q-factor models in explaining THSI-listed stock returns, demonstrating its relevance in Thai market analysis.

This research applies machine learning to predict sustainability ratings, focusing on overall ESG and individual pillar scores: environment (E score), social (S score), and governance (G score). These scores are essential for investors, regulators, and companies assessing corporate sustainability performance. Using financial data from 2020-2022, models like multiple linear regression, decision trees, random forests, and artificial neural networks were evaluated. Random forests performed best, with an R2 of 0.5089 for the ESG score and lower error rates (MAE: 0.0909, RMSE: 0.1140, MAPE: 22.71%). Company size was a key predictor of the overall ESG score, especially for environmental and social scores. These findings highlight random forests as a reliable tool for ESG assessment, aiding investment and governance evaluations.

details here

details here