Conference Agenda

In the last decade, stakeholders have been pressuring firms to take more initiative with their sustainability agendas, leading to a booming demand for firms’ non-financial information disclosure. Corporate Social Responsibility (CSR) performance is increasingly used to evaluate organizations’ management quality, identify their exposure to risk, and assess their ability to leverage business opportunities. This leads to the question: Does CSR affect corporate operation efficiency. This study analyzes Singaporean companies’ sustainability reports for 2015 to 2021 and their relation with corporate operational efficiency. Three research questions are addressed:

RQ1: What are the disclosure characteristics including topic components, sentiment and readability in sustainability reports during the period of study?

RQ2: How do those topics identified in RQ1 relate to corporate operation efficiency score computed using Data Envelopment Analysis (DEA)?

RQ3: How do moderating factors including product market competition, ESG risk scores and board gender diversity impact findings from RQ2?

The sample consists of 206 sustainability reports from companies included in the Stratis Time Index, the benchmark index for Singapore’s stock market. Several NLP methods are employed – Tokenization and Latent Dirichlet Allocation (LDA), readability assessment, and sentiment analysis. The second objective of this study is to determine the relation between sustainability disclosures and DEA operation efficiency scores. After controlling for Return on Assets (ROA), Market Value (MV), firm age, industry and fixed year effect, the findings suggest a positive association. Furthermore, analyses are conducted to study the moderating effects of external monitoring and corporate governance mechanism on the magnitude of the association.

New regulatory frameworks such as LkSG (Germany) or CSDDD (EU), increase the need for tracking and build-up of KPI frameworks in companies with physical Supply Chains to disclose the level of social sustainability (e.g., child or forced labor).

Due to lack of KPI frameworks and methodologies provided from regulatory side (based on legal document analysis), the methods in practical use by companies subject to strict legal requirements, were collected and analysed in a longitudinal design by a self-programmed text mining tool on German public annual reports.

The annual report results were combined with a logit regression approach to uncover patterns between disclosure methodologies in use and social dimensions of ESG Ratings as a measure of success, to give operations research (OR) scholars and practitioners a guidance on current dynamics and trends in an emerging field reacting to new legal requirements ensuring basic human rights in supply chains.

We focus on exploring the adaptation and success of certain disclosure methods for Social Sustainability, shortcomings of the existing studies, and providing insights for practitioners as well as future directions for researchers.

The study contributes to the OR field by offering an empiric view on methodologies and KPI frameworks, as well as industry specific norms to optimize a firm’s operations for success in the social sustainability dimension.

Textile sector is one of the most polluting and waste generating sectors. Pertinent efforts have been made to reduce its environmental impact and bring sustainability to its operations. Government is continuously promoting patent granting and filing in this sector to develop innovative solutions for sustainability. Hence, there is continuous increase in the number of patents published in the last decade. These patents are commonly described in text formats that can be analysed by Artificial Intelligence tools such as Natural Language Processing. In this study, an analysis of patents in textile field is conducted to extract keywords from patent abstracts using algorithms such as TF-IDF and TextRank. Identified keywords are used for weight computation and string matching. Weight computation is done by calculating the frequency of each term using TF-IDF and Text Rank whereas string matching is done using Levenshtein-distance to reduce repetition of terms. Finally, network analysis is performed to understand the relations between keywords and to find the most influential technologies for sustainability in textile sector.