FinTech—from blockchain to AI—is remaking capital markets, improving liquidity and valuation efficiency but also concentrating information and increasing algorithmic opacity; these forces can speed markets yet complicate sustainability goals and raise crash risk.

Main Finding

The paper develops a unified conceptual framework linking digital transformation (FinTech innovation) to stock exchange sustainability (ESG outcomes) and share valuation efficiency. It synthesizes global empirical evidence to map technological inputs (blockchain, AI/ML, big data, digital trading platforms, ESG reporting tech) through transmission mechanisms (transparency, liquidity, information processing, governance) to capital market outcomes (valuation, liquidity, information asymmetry, crash risk). The framework highlights both efficiency gains and emergent risks (informational concentration, algorithmic opacity, contextual heterogeneity) and emphasizes feedback loops and structural moderators (market development stage, institutional/regulatory environments).

Key Points

• Framework architecture: three layers — (1) Innovation Input (technologies), (2) Transmission Mechanisms (efficiency, sustainability, governance channels), (3) Capital Market Outcomes (valuation, liquidity, ESG performance) — with feedback loops and contextual moderators.
• Core technologies considered:
  • Blockchain & smart contracts: transparency, traceability, tokenization, faster settlement; enables ESG traceability and green finance but faces scalability and regulatory ambiguity.
  • AI/ML: predictive analytics, risk modelling, algorithmic trading, fraud detection; improves efficiency and risk management but raises model opacity, bias, and systemic risks.
  • Big data & AIoT: soft-data integration, real-time sensing for ESG, improved credit allocation; raises privacy and data-governance challenges.
  • Digital trading platforms/HFT: enhanced liquidity and price discovery, but can create competitive asymmetries and increase volatility/short-termism.
  • ESG reporting technologies: machine-readable disclosures can make sustainability signals value-relevant.
• Channels from FinTech to corporate sustainability: green/digital finance reallocating capital, green innovation stimulation, organizational process innovation, supply-chain finance reducing Scope 3 emissions.
• Non-linear and context-dependent effects: evidence of U-shaped or diminishing marginal returns for digital finance on sustainability; outcomes differ across developed, emerging and frontier markets; digital infrastructure quality and regulation moderate effects.
• Risks treated as endogenous: framework integrates both benefits (efficiency, inclusion, improved disclosure) and risks (informational concentration, algorithmic opacity, potential for crashes, inequitable access).
• Scope limits: conceptual (not predictive), focused on formal/regulated capital markets (excludes DeFi/informal markets), mid-level theory to guide empirical testing.

Data & Methods

• Method: comprehensive, cross-regional literature review and conceptual synthesis. The paper aggregates empirical and theoretical studies from established markets (Europe, US), fast-growing Asian markets, and emerging African markets to build the causal mapping and propositions.
• Outputs: a multi-figure conceptual model (Figures 1–6 and others referenced) that:
  • identifies mechanisms by technology,
  • specifies structural moderators (market stage, regulatory regime, ESG infrastructure),
  • outlines feedback loops between market outcomes and technology investment.
• Evidence base: secondary sources and systematic reviews (papers and studies cited from ~2018–2025), cross-country empirical findings (banking and capital market studies), and technology-impact case syntheses. No original primary data or quantitative estimation in this paper.

Implications for AI Economics

Practical and research implications relevant to the economics of AI in capital markets:

• Roles played by AI in markets

  • Information processing: AI lowers information frictions by ingesting alternative/soft data, improving forecasts and credit allocation.
  • Market microstructure: AI-powered trading and HFT reshape liquidity provision, price discovery, and short-term volatility.
  • ESG valuation: AI enables automated, machine-readable ESG reporting and extraction of value-relevant sustainability signals from non-traditional data.
  • Risk management: AI improves risk modelling and fraud detection but introduces model opacity and correlated algorithmic behavior.

• Economic risks and distributive effects

  • Concentration of advantage: firms with superior AI/data infrastructure obtain persistent informational/latency advantages (raises market power, entry barriers).
  • Algorithmic opacity and bias: opaque models can produce systematic mispricing, regulatory arbitrage, and fairness concerns (credit, inclusion).
  • Systemic/market-stability risk: correlated AI strategies can amplify tail events and crash risk; algorithmic “herding” and potential tacit collusion.
  • Heterogeneous impacts across markets: effects differ by market maturity, data availability, and regulatory quality—AI may widen global inequalities in capital access and valuation efficiency.

• Policy, regulation, and market design suggestions

  • Mandate machine-readable, standardized ESG disclosures to enable AI-based comparability and reduce greenwashing.
  • Require model governance, explainability standards, and validation/auditing for AI systems that materially affect market prices or investor protection.
  • Monitor concentration metrics (e.g., HHI for algorithmic liquidity providers) and design market rules to mitigate asymmetric latency advantages.
  • Strengthen data-governance, privacy, and consent regimes, and ensure public-sector data infrastructure in emerging markets to reduce uneven AI benefits.
  • Explore market safeguards for algorithmic trading (circuit breakers, throttles, kill-switches) and routine stress-testing of algorithmic strategies.

• Research agenda / testable hypotheses for AI economics

  • Measure and identify causal effects:
    • Hypothesis A: Firm-level AI adoption (IT/R&D spend, AI patents, cloud compute usage) increases intraday liquidity but also raises intraday volatility; test with firm-time panel and microstructure metrics (bid-ask spread, depth, signed order flow).
    • Hypothesis B: Adoption of AI-based ESG analytics reduces cross-sectional mispricing of sustainability-exposed firms and lowers dispersion in ESG-adjusted returns; test with event studies around AI-ESG tool rollouts and difference-in-differences across firms/exchanges.
    • Hypothesis C: Higher concentration of algorithmic trading intensifies crash risk (tail-dependence); test using systemic risk measures, tail correlations, and changes in liquidity during stress episodes.
    • Hypothesis D: AI-driven disclosure increases market valuation of green projects in markets with standardized, machine-readable ESG reporting but has muted effects where ESG infrastructure is weak (interaction with institutional quality).
  • Measurement strategies:
    • Proxy AI adoption with IT expenditure composition, cloud vendor activity, job postings for ML roles, AI-related patents, and partnership announcements.
    • Quantify algorithmic concentration via trading activity shares, HFT participation rates, and latency metrics.
    • Assess ESG signal quality via disclosure completeness, machine-readability scores, and cross-provider score dispersion.
  • Methodologies:
    • Combine microstructure tick data, firm financials, ESG disclosure filings, and regulatory event windows.
    • Use IV, DiD, regression discontinuity around regulatory or platform changes, and structural models of information acquisition to identify causal pathways.
    • Cross-country panel studies exploiting heterogeneity in digital infrastructure and regulation to uncover moderating effects.

• Measurement and valuation challenges for AI-driven assets

  • Difficulty valuing scalable digital business models and intangible AI assets; standard accounting lags underlying value creation from AI capabilities.
  • Tokenization and digital securities raise questions for liquidity measurement, market microstructure, and taxation—need economic models to price tokenized equity/liability structures.

Limitations noted by the paper that affect AI-economics work - Conceptual (not predictive) — requires empirical validation and parameter estimation. - Focus on regulated/formal markets — DeFi and crypto-native mechanisms are outside scope and require separate analysis. - Context dependence — empirical designs must account for cross-jurisdictional institutional heterogeneity.

Overall, the framework provides a structured roadmap for AI economics research in capital markets: identify specific AI technologies and mechanisms, measure adoption and market outcomes, exploit cross-market heterogeneity for identification, and evaluate regulatory interventions aimed at maximizing efficiency gains while mitigating concentration, transparency, and stability risks.