A1530
Title: Economic narrative processing in the case of climate change
Authors: Hiroki Sakaji - The University of Tokyo (Japan)
Noriyasu Kaneda - Bank of Japan (Japan) [presenting]
Abstract: Important economic narratives are extracted from newspaper articles using a causal chain method and BERT, a deep learning-based language model that performs dependency parsing of economic contexts. Economic narratives are important causes of major economic events, and then they could provide clues for understanding people's beliefs and expectations. The novel framework can construct indices of economic narratives for any financial and economic issues and visualize a graphical linkage of subtopics. As a case study, we apply this method to describe a narrative for climate change. The result suggests that governments had major roles in discussing international frameworks and environmental regulations for carbon neutrality in the 2000s. Also, firms and green investors seemed to react to the progress of the climate policy debate and have tackled long-term risk management and investment in new businesses in recent years. The climate narrative could be useful for the analysis of spillover of policy announcements, expectation formation and behavioral changes in microeconomic entities. For future work, we need to test empirically whether the novel method could be an effective approach to dissecting people's beliefs in climate policy, market bubbles or inflation and embodying narrative economics.
