Opportunity or opportunism? Blockchain technology adoption and corporate default risk

The study examines whether blockchain technology adoption by firms reduces or increases corporate default risk, and through which mechanisms. Motivated by widespread attention and hype surrounding blockchain disclosures and mixed views about its maturity and value, the paper contrasts two perspectives: the technology acceleration perspective (blockchain improves efficiency, transparency, and reduces information asymmetry) and the managerial opportunism perspective (managers strategically disclose blockchain involvement to mislead investors, manipulate perceptions, and pursue self-interest). The research questions are: Does blockchain adoption increase or decrease corporate default risk? What mechanisms drive the effect in the Chinese capital market?

Institutional background highlights rapid growth and policy support for blockchain applications globally and in China. Two competing hypotheses emerge. Technology acceleration hypothesis: blockchain enhances data processing, transparency, creditworthiness, supply chain trust, and governance via consensus and smart contracts, thereby lowering default risk (e.g., Chen et al., 2019; Cheng et al., 2019; Xu and Zou, 2021). Evidence includes positive market reactions to blockchain-related disclosures and patents, and potential reductions in information asymmetry and trade credit risk. Managerial opportunism perspective: firms may strategically disclose blockchain adoption to inflate market value, time financing, and manage earnings, with rating agencies underpricing risk; prior studies document hype, manipulation, and increased default probability in speculative contexts (e.g., Cheng et al., 2019; Cioroianu et al., 2021; Griffin and Shams, 2020; Gandal et al., 2018; Cathcart et al., 2020). The literature on China’s market is relatively sparse, motivating this study.

Data and sample: Chinese A-share listed companies from 2001–2021. Initial 42,552 firm-year observations were filtered by removing those with missing data, financial firms and ST/PT firms, resulting in a final sample of 22,567 firm-years. Variables sourced from the China Research Data Services Platform (MD&A texts) and CSMAR. Data are winsorized at 1% and standard errors clustered at the firm level; industry and year fixed effects included. Blockchain adoption identification: Constructed an objective blockchain dictionary using word2vec trained on 49,654 annual reports (2001–2021). Starting from seed term “blockchain,” semantically similar terms (22 total; e.g., smart contract, distributed ledger) were selected via cosine similarity. Bag-of-Words on MD&A texts identified presence of blockchain-related terms. A binary indicator Blockchain equals 1 if terms appear in MD&A in year t, 0 otherwise. An alternative intensity proxy PBTV equals the ratio of blockchain-related words to total MD&A words. Default risk measurement: Expected default probability (EDP) based on Bharath and Shumway (2008). Compute distance-to-default (DD) using firm market equity, book debt, lagged annual return, asset volatility (derived from stock return volatility), and time horizon T=1. EDP is Normal(-DD). Alternative default proxies include: (i) log count of firm violations, (ii) a default dummy, and (iii) Altman-style Z-score (Chinese adaptation). Controls: Firm size (log assets), leverage, cash holdings, ROA, growth, SOE, fixed assets ratio, firm age (log), CEO duality, largest shareholder ownership (Hold1); plus year and industry fixed effects. Empirical strategy: Baseline OLS regressions of EDP on Blockchain and controls with year/industry FE. Mechanism tests: (a) managerial self-interest via executive shareholding changes (Change), reductions (Sell), increases (Buy) interacted with Blockchain; (b) power consolidation via CEO duality (Dual) and internal control environment (Internal_Control) interacted with Blockchain. Additional outcome analyses for capital structure: total debt ratio (DT), long-term debt (LT), short-term debt (ST). Investment efficiency using a Richardson (2006)-style framework: Efficiency, Overinvest, Underinvest. Endogeneity and robustness: Difference-in-differences (DID) around first blockchain adoption, event-study dynamics, propensity score matching (PSM) and PSM-DID. Instrumental variables (2SLS) using the interaction of historical city post office density in 1984 and lagged national internet user base (broadband users; log-transformed) as instruments for Blockchain, following Huang et al. (2019) and Yuan et al. (2021). Subsample tests around 2008 (emergence of blockchain) with and without firm fixed effects. Placebo tests with random treatment assignment repeated 500 times.

• Descriptive: 31.6% of firm-years indicate blockchain adoption (mean Blockchain=0.316). Mean EDP is 0.016 (1.62%).
• Main effect: Blockchain adoption increases expected default probability. In baseline regressions, Blockchain coefficient ≈ 0.004, significant at 1%. Size and leverage are positively associated with default risk; ROA, growth, and largest shareholder ownership are negatively associated; SOE is positively associated.
• Mechanisms—managerial self-interest: Interaction ChangeBlockchain ≈ 0.008 (5%); SellBlockchain ≈ 0.008 (1%); Buy*Blockchain not significant. Managers appear to profit primarily by selling shares around blockchain adoption rather than buying.
• Mechanisms—power consolidation: DualBlockchain ≈ 0.009 (10%), indicating higher default risk where CEO also chairs the board. Internal_ControlBlockchain ≈ 0.039 (1%), suggesting weaker internal control environments amplify the risk-increasing effect (as coded, absence of an internal control evaluation report indicates poorer control).
• Capital structure: Blockchain is associated with higher total debt (DT) ≈ 0.003 (10%) and higher short-term debt (ST) ≈ 0.005 (1%), but lower long-term debt (LT) ≈ -0.002 (5%). This points to short-termism in financing post-adoption.
• Investment efficiency: Blockchain relates to lower investment efficiency driven by overinvestment. Efficiency ≈ 0.002 (10%), Overinvest ≈ 0.003 (10%), Underinvest not significant; interpretation in text indicates distorted efficiency primarily via overinvestment.
• Alternative measures: Using PBTV (textual intensity), the coefficient on EDP ≈ 1.344 (1%), confirming the baseline. Alternative default proxies: Blockchain positively relates to firm violation count and dummy, and negatively to Z-score (≈ -0.003, 10%), consistent with increased default risk.
• Causal inference: DID ≈ 0.003 (5%) shows an increase in default risk after adoption; event study shows insignificant pre-trends and significant effects from the current to several post periods (post3 highly significant). PSM and PSM-DID also yield positive and significant effects (≈ 0.004 each). IV results: first-stage instruments are relevant; second-stage Blockchain coefficient on EDP ≈ 0.073 (5%), reinforcing a positive causal effect.
• Subsamples: Effects are stronger and significant after 2008; full sample with firm FE becomes weaker, indicating the phenomenon intensifies in the era when blockchain became salient.

Findings support the managerial opportunism perspective: firms that disclose blockchain adoption experience higher expected default risk, consistent with managers leveraging hype to pursue short-term gains and mask performance. The effect is amplified when executive power is consolidated (CEO duality, weak internal controls), and when managerial self-interest is salient (executive share sales). Real effects manifest in increased overall and short-term debt and overinvestment, consistent with short-termism and risk-taking post-adoption. The combination of baseline, DID/event-study, PSM/PSM-DID, alternative measures, and IV analysis strengthens causal interpretation. These results nuance the narrative around blockchain’s benefits by documenting a “dark side” in capital market behavior, particularly in contexts with weaker governance and oversight.

The paper shows that blockchain adoption by Chinese listed firms is associated with higher corporate default risk, supporting the managerial opportunism hypothesis over the technology acceleration hypothesis in this setting. The risk increase is stronger under CEO power consolidation (duality, poor internal controls) and aligns with managerial self-interest behaviors (executive share sales). Blockchain adoption corresponds to greater short-term debt financing and overinvestment, indicating real effects beyond disclosure. The results are robust across alternative measurements, quasi-experimental designs (DID/event study, PSM-DID), and instrumental variables. Contributions include documenting the adverse governance-related consequences of blockchain adoption in an emerging market, informing investors, regulators, and firms about potential misuse of technology narratives. Future research could examine cross-country generalizability, refine measurement of true blockchain implementation versus disclosure, and explore governance mechanisms that mitigate opportunistic use of technology announcements.

• Measurement of blockchain adoption relies on textual analysis of MD&A; it may miss non-disclosed implementations or misclassify hype. Future work could incorporate investment data, project-level information, and machine learning to better distinguish real adoption from mere signaling.
• The literature and theoretical frameworks on blockchain’s impact on corporate behavior are still developing; more work is needed to validate and extend mechanisms across settings.
• Generalizability to developed markets with stricter legal and governance regimes is uncertain; cross-country studies are needed to assess external validity.