Constructing a Blockchain Disclosure Database for Supply Chain Finance Analytics: Design, Validation, and Firm-Level Applications
Main article
Abstract
Empirical research on blockchain technology in corporate governance has expanded rapidly, yet most studies still rely on hand-coded indicators that are difficult to audit and reproduce. This article responds to that gap by constructing the Blockchain Disclosure Database (BDD), a firm-year panel that combines text-mined disclosures from Chinese A-share annual reports with structured supply chain finance variables drawn from CSMAR and Wind. The article describes the database design, the validation hierarchy used to assess construct and predictive validity, and three firm-level applications that show how the database supports analytical workflows in supply chain finance research. The constructed database covers 28,479 firm-year observations across ten sectors during 2015 to 2023 and reaches a final blockchain disclosure rate of approximately 78 percent in the most recent years. Validation tests indicate strong agreement with manual audit samples, sector-level coverage that is consistent with industry technology trends, and predictive validity for accounting information disclosure quality both directly and through a supply chain finance mediation channel. Threshold tests confirm that the marginal effect of blockchain disclosure on disclosure quality varies non-linearly with firm size, with larger firms gaining the most. The paper argues that database design choices, not just econometric techniques, shape the inferences drawn in this research field. The BDD is therefore presented as a transparent and reusable empirical infrastructure for accounting, finance, and supply chain analytics.
