Edoardo Crocco, Laura Broccardo, Hind Alofaysan, Reeti Agarwal
{"title":"碳密集型行业的可持续性报告:跨行业机器学习方法的启示","authors":"Edoardo Crocco, Laura Broccardo, Hind Alofaysan, Reeti Agarwal","doi":"10.1002/bse.3850","DOIUrl":null,"url":null,"abstract":"<p>Due to climate change concerns, academics and practitioners focus more on environmental management and sustainability. Accounting researchers have focused on corporate environmental disclosure and sustainability reporting in response to stakeholder demand for openness and accountability. Thus, scholarly studies on sustainability reporting have gained momentum with the frequent use of qualitative text analysis to assess company disclosures' completeness and quality. However, sustainability reporting research has major limitations wherein past studies have focused on certain sectors or qualitative content analysis. Coherently with the abovementioned gap, our study intends to examine sustainability reports of companies in agriculture, conventional energy, heavy industry and manufacturing, transport and automotive, and construction, the highly carbon-intensive industries most vulnerable to physical climate damage and net-zero transition risk. In doing so, the goal of the present research is to investigate sustainability reporting practice on a larger, cross-sectoral scale by using automated, machine learning-powered text analysis methods to complement and strengthen qualitative research results that scholars have previously obtained. The latent Dirichlet allocation topic modelling technique has been used to examine companies' sustainability efforts and identify industry-specific subtopics based on quantitative distribution. The originality of our analysis lies in determining how companies prioritise issues in sustainability reports. By comparing reports from different industries, we also identify sector-specific patterns and how organisations in highly carbon-intensive industries that are most exposed to physical climate damage and net-zero transition risk prioritise certain themes over others, as well as identifying what type of content is overall more prominently featured in reports, regardless of the industry. Our study adds to sustainability reporting literature by investigating a previously unstudied sample of sectors. Moreover, our study informs practitioners of existing sustainability reporting procedures. The subject model and a cross-industry view can advise policymakers and industry of which themes are under-disclosed and what industry-specific rules may be desirable to suit sector-specific needs.</p>","PeriodicalId":9518,"journal":{"name":"Business Strategy and The Environment","volume":"33 7","pages":"7201-7215"},"PeriodicalIF":12.5000,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Sustainability reporting in carbon-intensive industries: Insights from a cross-sector machine learning approach\",\"authors\":\"Edoardo Crocco, Laura Broccardo, Hind Alofaysan, Reeti Agarwal\",\"doi\":\"10.1002/bse.3850\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Due to climate change concerns, academics and practitioners focus more on environmental management and sustainability. Accounting researchers have focused on corporate environmental disclosure and sustainability reporting in response to stakeholder demand for openness and accountability. Thus, scholarly studies on sustainability reporting have gained momentum with the frequent use of qualitative text analysis to assess company disclosures' completeness and quality. However, sustainability reporting research has major limitations wherein past studies have focused on certain sectors or qualitative content analysis. Coherently with the abovementioned gap, our study intends to examine sustainability reports of companies in agriculture, conventional energy, heavy industry and manufacturing, transport and automotive, and construction, the highly carbon-intensive industries most vulnerable to physical climate damage and net-zero transition risk. In doing so, the goal of the present research is to investigate sustainability reporting practice on a larger, cross-sectoral scale by using automated, machine learning-powered text analysis methods to complement and strengthen qualitative research results that scholars have previously obtained. The latent Dirichlet allocation topic modelling technique has been used to examine companies' sustainability efforts and identify industry-specific subtopics based on quantitative distribution. The originality of our analysis lies in determining how companies prioritise issues in sustainability reports. By comparing reports from different industries, we also identify sector-specific patterns and how organisations in highly carbon-intensive industries that are most exposed to physical climate damage and net-zero transition risk prioritise certain themes over others, as well as identifying what type of content is overall more prominently featured in reports, regardless of the industry. Our study adds to sustainability reporting literature by investigating a previously unstudied sample of sectors. Moreover, our study informs practitioners of existing sustainability reporting procedures. The subject model and a cross-industry view can advise policymakers and industry of which themes are under-disclosed and what industry-specific rules may be desirable to suit sector-specific needs.</p>\",\"PeriodicalId\":9518,\"journal\":{\"name\":\"Business Strategy and The Environment\",\"volume\":\"33 7\",\"pages\":\"7201-7215\"},\"PeriodicalIF\":12.5000,\"publicationDate\":\"2024-07-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Business Strategy and The Environment\",\"FirstCategoryId\":\"91\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/bse.3850\",\"RegionNum\":1,\"RegionCategory\":\"管理学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"BUSINESS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Business Strategy and The Environment","FirstCategoryId":"91","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/bse.3850","RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BUSINESS","Score":null,"Total":0}
Sustainability reporting in carbon-intensive industries: Insights from a cross-sector machine learning approach
Due to climate change concerns, academics and practitioners focus more on environmental management and sustainability. Accounting researchers have focused on corporate environmental disclosure and sustainability reporting in response to stakeholder demand for openness and accountability. Thus, scholarly studies on sustainability reporting have gained momentum with the frequent use of qualitative text analysis to assess company disclosures' completeness and quality. However, sustainability reporting research has major limitations wherein past studies have focused on certain sectors or qualitative content analysis. Coherently with the abovementioned gap, our study intends to examine sustainability reports of companies in agriculture, conventional energy, heavy industry and manufacturing, transport and automotive, and construction, the highly carbon-intensive industries most vulnerable to physical climate damage and net-zero transition risk. In doing so, the goal of the present research is to investigate sustainability reporting practice on a larger, cross-sectoral scale by using automated, machine learning-powered text analysis methods to complement and strengthen qualitative research results that scholars have previously obtained. The latent Dirichlet allocation topic modelling technique has been used to examine companies' sustainability efforts and identify industry-specific subtopics based on quantitative distribution. The originality of our analysis lies in determining how companies prioritise issues in sustainability reports. By comparing reports from different industries, we also identify sector-specific patterns and how organisations in highly carbon-intensive industries that are most exposed to physical climate damage and net-zero transition risk prioritise certain themes over others, as well as identifying what type of content is overall more prominently featured in reports, regardless of the industry. Our study adds to sustainability reporting literature by investigating a previously unstudied sample of sectors. Moreover, our study informs practitioners of existing sustainability reporting procedures. The subject model and a cross-industry view can advise policymakers and industry of which themes are under-disclosed and what industry-specific rules may be desirable to suit sector-specific needs.
期刊介绍:
Business Strategy and the Environment (BSE) is a leading academic journal focused on business strategies for improving the natural environment. It publishes peer-reviewed research on various topics such as systems and standards, environmental performance, disclosure, eco-innovation, corporate environmental management tools, organizations and management, supply chains, circular economy, governance, green finance, industry sectors, and responses to climate change and other contemporary environmental issues. The journal aims to provide original contributions that enhance the understanding of sustainability in business. Its target audience includes academics, practitioners, business managers, and consultants. However, BSE does not accept papers on corporate social responsibility (CSR), as this topic is covered by its sibling journal Corporate Social Responsibility and Environmental Management. The journal is indexed in several databases and collections such as ABI/INFORM Collection, Agricultural & Environmental Science Database, BIOBASE, Emerald Management Reviews, GeoArchive, Environment Index, GEOBASE, INSPEC, Technology Collection, and Web of Science.