Fintech expansion lifts productivity at listed Chinese manufacturers by improving supply-chain finance and spurring innovation, but government subsidies blunt the gains; effects are largest in high‑tech industries and richer regions.
Research on the Impact of Financial Technology on the Total Factor Productivity of Manufacturing Enterprises
Using a panel of Chinese A-share manufacturing firms (2015–2023), the study finds that fintech development raises firm-level total factor productivity mainly through improved supply-chain finance and innovation, with government subsidies weakening this effect and stronger impacts in high-tech sectors and wealthier regions.
Driven by digital technologies including big data and artificial intelligence, financial technology (fintech) has emerged as a key engine for boosting the total factor productivity of manufacturing firms, and its underlying impact channels have been widely discussed in academic research. Based on a sample of Chinese A-share listed manufacturing enterprises covering the period 2015–2023, this study empirically confirms that fintech development can significantly enhance corporate total factor productivity primarily through the channels of supply chain finance and innovation effects. Meanwhile, government subsidies exert a negative moderating influence on this relationship. Further heterogeneity tests indicate that the productivity-enhancing effect of fintech is more pronounced in high-tech industries and regions with higher levels of economic development.
Summary
Main Finding
Fintech development significantly increases total factor productivity (TFP) of Chinese manufacturing firms (A‑share listed, 2015–2023). The effect operates primarily through two channels — supply‑chain finance (liquidity/asset activation) and innovation (more patenting) — while government subsidies negatively moderate the fintech → TFP relationship. The productivity gain from fintech is stronger for high‑tech firms and in economically advanced regions.
Key Points
- Hypotheses tested
- H1: Fintech positively affects manufacturing firm TFP. (Supported)
- H2: Fintech enhances TFP via supply‑chain finance. (Supported)
- H3: Fintech enhances TFP via innovation (patent activity). (Supported)
- H4: Government subsidies negatively moderate the fintech → TFP effect. (Supported)
- Magnitudes (selected results)
- Baseline fintech coefficient on TFP (with controls, firm & year fixed effects): ≈ 0.271 (significant at 1%).
- Fintech → ln(Scf) (supply‑chain finance): coefficient ≈ 0.420 (≈10% level).
- ln(Scf) → TFP: coefficient ≈ 0.115 (1%).
- Fintech → Patent (log patent apps+1): ≈ 0.482 (5%).
- Patent → TFP: ≈ 0.061 (1%).
- Interaction FT × Gov: ≈ −0.057 (significant at 5%), indicating subsidies weaken the fintech effect.
- Controls: standard firm controls included (ROA, leverage, age, fixed asset ratio, growth, cash flow, board size, largest shareholder share, regional lnGDP). High adjusted R2 (~0.92).
- Heterogeneity: stronger fintech → TFP effect in high‑tech sectors and in regions with higher economic development.
Data & Methods
- Sample: Chinese A‑share listed manufacturing firms, 2015–2023. Observations reported N ≈ 12,661 (panels with firm and time fixed effects).
- Dependent variable: firm TFP estimated by the Levinsohn‑Petrin (LP) method (log TFP = log output − estimated input elasticities × inputs).
- Output: operating income.
- Inputs: capital = original value of fixed assets; labor = number of employees; materials intermediate measure M constructed from costs minus depreciation etc.
- Main explanatory variable: regional fintech development index following Li et al. (operationalized as a regional fintech metric; details per cited methodology).
- Mediators:
- Supply‑chain finance (Scf) = ln(short‑term loans + accounts payable + 1).
- Innovation (Patent) = ln(number of patent applications + 1).
- Moderator: Government support intensity (Gov) = ratio of government funds to investment in scientific & technological innovation.
- Econometric approach:
- Panel regressions with firm and year fixed effects.
- Mediation analysis via two-step regressions (Fintech → mediator; mediator → TFP controlling for fintech).
- Moderation tested via interaction term Fintech × Gov.
- Key robustness/diagnostics: paper reports significance levels, t‑stats, and high adjusted R2; no detailed instrumental variable or dynamic panel treatment reported in the main summary.
Implications for AI Economics
- Mechanisms linking AI‑enabled fintech to productivity
- AI and big‑data components of fintech (credit scoring, risk models, automated contracts) increase liquidity allocation efficiency (supply‑chain finance) and reduce frictions for innovative firms, raising firm‑level TFP.
- AI tools also support firm digitalization that converts data into productive capital, amplifying R&D returns and patenting activity.
- Policy implications
- Targeted fintech support and regulatory frameworks that encourage AI‑driven financial services can improve manufacturing productivity, especially in tech‑intensive sectors and wealthier regions.
- Broad or poorly targeted government subsidies risk crowding out private innovation incentives; subsidy design should include transparency, monitoring (potentially digitalized), and conditionality to support R&D rather than routine operating costs.
- Promote integration of fintech with industrial policy (e.g., supply‑chain finance products tailored to SME suppliers) while guarding against rent seeking.
- Directions for AI economics research
- Causal identification: use instruments or quasi‑experimental designs to address endogeneity (e.g., staggered fintech policy rollouts, exogenous shocks to digital infrastructure).
- Disaggregate AI components: estimate separate impacts of AI modules (ML credit scoring, NLP for compliance, smart contracts/blockchain) on finance channels and productivity.
- Distributional and labor effects: examine how fintech‑AI affects employment composition, wages, and skill demand within manufacturing.
- Firm heterogeneity: explore size, ownership (state vs. private), and balance sheet constraints as moderators of fintech effects.
- Cross‑country comparisons: test whether the documented channels hold in economies with different financial structures and regulatory regimes.
- Dynamic outcomes: study medium/long‑run productivity dynamics, diffusion of fintech across supply chains, and possible general equilibrium effects.
Limitations to note when using these results: measurement of the fintech index is region‑level and may mask firm‑level fintech adoption; potential reverse causality (more productive regions attract fintech); analysis restricted to listed manufacturing firms in China, which may limit generalizability.
Assessment
Paper Typequasi_experimental
Evidence Strengthmedium — The study uses longitudinal firm-level data, fixed effects, mediation and heterogeneity analyses which strengthen associative inference, but it appears to lack a clearly exogenous source of variation (e.g., an instrument or quasi-random policy shock) to fully rule out endogeneity, reverse causality, and omitted variable bias.
Methods Rigormedium — Methodological choices (panel regressions, controls, mediation and heterogeneity analyses) are appropriate and standard for firm-level productivity work; however, the absence of a clearly exogenous identification strategy, potential measurement issues for fintech and TFP, and limited detail on robustness to dynamic endogeneity reduce overall rigor.
SampleFirm-year panel of Chinese A-share listed manufacturing enterprises covering 2015–2023; outcome is firm total factor productivity constructed from financial statement data; key predictor is a measure/index of fintech development (likely at regional or industry level); analyses control for firm characteristics and include year and firm fixed effects.
Themesproductivity innovation adoption
IdentificationUses firm-level panel regressions exploiting cross-sectional and time variation in fintech development and firm outcomes, with firm and year fixed effects, firm controls, mediation tests for supply-chain finance and innovation channels, and heterogeneity/robustness checks; no clear randomized assignment or natural experiment/instrument reported, so causal claims rely on observational variation and controls.
GeneralizabilityLimited to publicly listed manufacturing firms (excludes SMEs and non-listed firms), China-specific institutional and fintech-development context may not generalize to other countries, Covers 2015–2023, a period of rapid fintech growth in China—effects may differ in earlier/later periods, Fintech as measured likely bundles many technologies (big data, AI, platforms), so results may not isolate pure AI effects, TFP measurement and fintech index construction choices may affect external validity
Claims (5)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| Fintech development can significantly enhance corporate total factor productivity for Chinese A-share listed manufacturing firms. Firm Productivity | positive | high | corporate total factor productivity |
0.48
|
| The positive effect of fintech on corporate total factor productivity operates primarily through the channels of supply chain finance and innovation effects. Firm Productivity | positive | high | corporate total factor productivity (through supply chain finance and innovation channels) |
0.48
|
| Government subsidies exert a negative moderating influence on the relationship between fintech development and corporate total factor productivity. Firm Productivity | negative | high | corporate total factor productivity (moderated by government subsidies) |
0.48
|
| The productivity-enhancing effect of fintech is more pronounced in high-tech industries. Firm Productivity | positive | high | corporate total factor productivity (heterogeneous effect by industry tech-intensity) |
0.48
|
| The productivity-enhancing effect of fintech is stronger in regions with higher levels of economic development. Firm Productivity | positive | high | corporate total factor productivity (heterogeneous effect by regional development) |
0.48
|