Unstructured ESG Data in Financial Modeling: Challenges and Opportunities in Emerging Markets

Environmental, Social and Governance (ESG) considerations have become central to investment analysis, creating a growing need for structured, comparable and usable ESG data. In many emerging markets, however, standardized ESG scores remain unavailable or cover only a limited fraction of listed companies. Much of the available ESG information exists instead in unstructured textual formats, including annual reports, sustainability disclosures, regulatory filings and corporate communications. This creates a methodological challenge, as traditional financial models cannot directly integrate qualitative and fragmented textual information.This paper examines the obstacles that make unstructured ESG data difficult to use in emerging market contexts, including data fragmentation, inconsistent reporting, multilingual documents and the risk of overstated sustainability narratives. It then discusses the role of Natural Language Processing (NLP) and related artificial intelligence techniques in transforming textual ESG disclosures into structured indicators. Building on this discussion, the paper proposes a five-phase conceptual framework that maps the transformation process from raw ESG documents to financial model inputs. Morocco is used as an illustrative context to show the institutional and data-related challenges faced by emerging markets. The paper contributes by offering a methodological blueprint for operationalizing unstructured ESG information and by connecting sustainable finance, computational text analysis and emerging market research.