China's supply-chain digitalization pilots shift firms' workforces toward higher-skilled employees by improving market visibility, brand trust and financing access; the effect is strongest in uncertain, competitive environments and in the eastern region.

How Artificial Intelligence Shapes the Human Capital Structure: Evidence from The Supply Chain Digitalization Pilots

Jiayi Fan · April 15, 2026 · Journal of innovation and development

openalex quasi_experimental medium evidence 7/10 relevance DOI Source PDF

Using a difference-in-differences design on Chinese A-share firms, the paper finds that supply-chain digitalization via pilot programs shifts firms' human capital toward higher-skilled workers, mediated by greater market attention, brand trust, eased financing constraints, and accumulation of digital intangible assets.

Against the backdrop of the in-depth advancement of the Digital China initiative and the accelerated restructuring of supply chain systems, digital technologies are profoundly reshaping the internal factor allocation structure of enterprises by remodeling the information flow, capital flow and collaboration modes among them. Based on the data of A-share listed companies on the Shanghai and Shenzhen Stock Exchanges from 2013 to 2022, this paper takes the pilots of supply chain innovation and application as a quasi-natural experiment and employs the difference-in-differences method to investigate the impact of supply chain digitalization on enterprises' human capital structure and its underlying mechanisms. The study finds that supply chain digitalization drives the optimization of the human capital structure. Mechanism analysis shows that supply chain digitalization enhances enterprises' capacity to absorb high-skilled labor through channels such as raising enterprises' market attention, boosting public trust in brands, alleviating financing constraints and promoting the accumulation of digital intangible assets. Heterogeneity analysis reveals that this facilitating effect is more pronounced in enterprises facing higher external environmental uncertainty, operating in more competitive industries and located in the eastern region of China. From the perspective of supply chain network collaboration and factor reallocation, this paper uncovers the micro-mechanism through which digital policies drive the upgrading of corporate human capital, and provides empirical evidence and practical insights for deepening supply chain digitalization construction and optimizing the talent structure of enterprises.

Summary

Main Finding

Supply-chain digitalization (exploiting China’s 2018 "Pilots of Supply Chain Innovation and Application") causally promotes the upgrading of corporate human capital structure in A‑share listed firms (2013–2022). Firms exposed to the pilots increase their capacity to absorb high‑skilled labor. The effect operates through greater external visibility (market attention and brand trust), eased financing constraints, and accumulation of digital intangible assets, and is stronger in firms facing higher environmental uncertainty, in more competitive industries, and in eastern China.

Key Points

Research question: Does supply‑chain digitalization — an institutional push for upstream–downstream digital collaboration — change firms’ internal human capital structure (toward more/high‑skilled labor)?
Identification strategy: The 2018 pilot program is treated as a quasi‑natural experiment; difference‑in‑differences (DID) is used to estimate the causal effect.
Main mechanisms identified:
- Increased market attention and external visibility reduce information asymmetry, lowering search/judgment costs for high‑skilled workers.
- Stronger brand/public trust increases stable employment expectations, making firms more attractive to skilled talent.
- Alleviation of financing constraints enables firms to invest in human capital and offer competitive compensation.
- Accumulation of digital intangible assets (e.g., data, platform capabilities) raises demand for and complementarities with skilled labor.
Heterogeneity: Larger effects where firms operate under greater external uncertainty, in highly competitive industries, and in more developed (eastern) regions.
Contribution: Shifts focus from aggregate employment effects of AI to micro‑level reallocation within firms’ human capital; highlights the role of supply‑chain network collaboration (not just internal digital transformation).

Data & Methods

Sample: Chinese A‑share listed companies on Shanghai and Shenzhen exchanges, 2013–2022.
Treatment: Inclusion in/coverage by the national supply‑chain digitalization pilots launched in 2018 (55 pilot cities, 266 pilot enterprises across sectors).
Empirical approach:
- Difference‑in‑differences framework comparing treated vs. control firms before/after the pilot rollout.
- Robustness checks (as reported): likely tests for parallel trends, placebo tests, and alternative specifications (paper reports robustness but specifics should be checked in the full text).
Outcomes and variables:
- Dependent variable(s): firm‑level measures of human capital structure (indicators of high‑skilled labor share / skill composition / human capital level — see paper for exact definitions).
- Mechanism variables: market attention, measures of brand/trust, financing constraints, indicators of digital intangible asset accumulation.
Further analyses: mechanism regressions and heterogeneity (by regional, industry, and firm‑environment characteristics).

Implications for AI Economics

Policy design: Digitalization policies that target supply‑chain collaboration can be an effective lever to reallocate labor toward higher‑skilled roles, beyond firm‑level digital adoption. Policymakers should consider supply‑chain‑level interventions when aiming to foster skill upgrading.
Complementarity view of AI: Results support the view that AI/digital technologies are skill‑biased and complement high‑skilled workers — especially when information frictions, financing constraints, and weak external signals are mitigated.
Role of institutions and networks: Institutional coordination (pilot programs, local support) and network effects across suppliers/customers materially affect firms’ ability to attract and retain skilled labor. Research and policy should account for upstream–downstream spillovers, not only within‑firm investments.
Targeting inequality and transition risks: Because effects concentrate in more developed regions and competitive industries, policymakers should prepare targeted training, mobility, and finance programs in lagging regions/industries to avoid widening spatial or sectoral skill gaps.
Research directions: Further work should (a) unpack which firm‑level human capital metrics change (wages, task content, occupation mix), (b) measure long‑run employment vs. compositional effects, and (c) examine interactions with labor‑market institutions (training, mobility, social insurance) to assess distributional consequences of AI-driven digitalization.

If you want, I can extract the exact variable definitions, econometric specifications, and robustness tests from the full paper (e.g., how human capital and the mechanism variables are constructed), or produce a one‑page slide summarizing these results for presentation.

Assessment

Paper Typequasi_experimental Evidence Strengthmedium — The DiD panel design over ten years and the exploration of multiple mechanisms give credible, moderately strong evidence for an association consistent with a causal effect; however, treatment (pilot) selection may be endogenous (non-random selection into pilots), parallel-trends and measurement assumptions may be violated, and results are limited to publicly listed Chinese firms, which weakens causal certainty and external validity. Methods Rigormedium — The authors employ a standard quasi-experimental DiD approach with mechanism and heterogeneity checks, which is appropriate and informative; but the rigor depends on untestable assumptions (exogeneity of pilot assignment, parallel trends), the quality of the digitalization exposure measure, and how robustly they address potential confounders (propensity, event studies, placebo tests, controlling for pre-trends). Without strong evidence that selection into pilots is as-good-as-random, methods are not top-tier causal. SampleFirm-year panel of A-share listed companies on the Shanghai and Shenzhen Stock Exchanges from 2013 to 2022; treatment defined by firms participating in supply chain innovation/application pilot programs; analyses appear to use corporate financials, human capital measures, measures of market attention and brand trust, financing constraint indicators, and digital intangible asset measures (sample restricted to publicly listed Chinese firms). Themeslabor_markets skills_training IdentificationTreated firms are those selected as pilots for supply chain innovation and application; the paper uses a difference-in-differences (DiD) design on a firm-year panel of A-share listed companies (2013–2022) to compare changes in treated firms before and after the pilot to changes in non-treated firms, plausibly controlling for time-invariant firm heterogeneity and common time effects (firm and year fixed effects and observable covariates). Mechanism tests (market attention, brand trust, financing constraints, digital intangible assets) and heterogeneity analyses supplement the DiD. GeneralizabilityLimited to publicly listed A-share firms in China (excludes private firms and SMEs), Context-specific to China's Digital China initiative and the particular pilot program design, Pilot participants may be non-random (larger or better-connected firms), limiting external validity, Findings reflect 2013–2022 period and may not generalize to later stages of AI-specific adoption, Supply-chain digitalization is broader than—and not identical to—AI adoption, so results may not map directly to AI-driven effects

Claims (9)

Claim	Direction	Confidence	Outcome	Details
Supply chain digitalization drives the optimization of the human capital structure of enterprises. Skill Acquisition	positive	high	human capital structure (optimization of workforce composition toward higher-skilled labor)	0.48
Supply chain digitalization enhances enterprises' capacity to absorb high-skilled labor by increasing firms' market attention. Skill Acquisition	positive	high	market attention (as a channel affecting absorption of high-skilled labor)	0.48
Supply chain digitalization enhances enterprises' capacity to absorb high-skilled labor by boosting public trust in brands. Skill Acquisition	positive	high	public/brand trust (as a channel affecting absorption of high-skilled labor)	0.48
Supply chain digitalization enhances enterprises' capacity to absorb high-skilled labor by alleviating financing constraints. Skill Acquisition	positive	high	financing constraints (reduction) as a mediating channel	0.48
Supply chain digitalization enhances enterprises' capacity to absorb high-skilled labor by promoting the accumulation of digital intangible assets. Skill Acquisition	positive	high	accumulation of digital intangible assets (as a channel affecting absorption of high-skilled labor)	0.48
The positive effect of supply chain digitalization on optimizing human capital structure is stronger for enterprises facing higher external environmental uncertainty. Skill Acquisition	positive	high	interaction effect between supply chain digitalization and external environmental uncertainty on human capital structure	0.48
The positive effect of supply chain digitalization on human capital structure is stronger for enterprises operating in more competitive industries. Skill Acquisition	positive	high	interaction effect between supply chain digitalization and industry competition on human capital structure	0.48
The positive effect of supply chain digitalization on human capital structure is stronger for enterprises located in the eastern region of China. Skill Acquisition	positive	high	regional heterogeneity (eastern vs. other regions) in the impact on human capital structure	0.48
This paper treats pilots of supply chain innovation and application as a quasi-natural experiment and employs a difference-in-differences method to identify causal effects of supply chain digitalization. Other	null_result	high	methodological identification strategy (use of pilot policy as quasi-natural experiment; DID estimation)	0.8