China’s digital economy has shifted jobs toward services and away from industry across provinces between 2013 and 2024, led more by firms digitally integrating production than by pure digital industries; the 2017 national digital strategy marked a turning point after which digital industrialization began to support manufacturing while industrial digitalization continued to propel servicization.

The impact of China's digital economy development on changes in the labor structure

Xinyan Luo · May 21, 2026 · Journal of Applied Economics and Policy Studies

openalex quasi_experimental medium evidence 8/10 relevance DOI Source PDF

Using Chinese provincial data from 2013–2024, the study finds that the digital economy accelerated servicization and deindustrialization of employment overall, with industrial digitalization exerting a stronger push toward services than digital industrialization, and with heterogeneous effects across urban/rural areas and application domains.

The digital economy has become a key force reshaping industrial structure and patterns of labor allocation. Against this backdrop, whether the digital economy has accelerated the trends of deindustrialization and servicization in the labor structure has emerged as an important issue requiring urgent clarification. Based on China's provincial panel data from 2013 to 2024, this study systematically examines the impact of the digital economy on changes in the labor structure from two dimensions: digital industrialization and industrial digitalization. The findings are as follows. First, the development of the digital economy generally promotes the servicization and deindustrialization of the labor structure, with the driving effect of industrial digitalization being stronger than that of digital industrialization. Second, the elevation of the "digital economy" to a national strategy in 2017 constituted a critical turning point. Thereafter, digital industrialization shifted toward promoting industrialization and restraining servicization, whereas industrial digitalization continued to strengthen servicization while suppressing industrialization. Third, the heterogeneity analysis shows that the urban digital economy exerts a stronger effect than the rural digital economy in promoting servicization and inhibiting industrialization. In addition, the impact of household-side digital economy applications is significantly greater than that of government and enterprise-side applications.

Summary

Main Finding

China’s digital-economy development (2013–2024, provincial panel) has on average accelerated servicization and reduced industrialization of the labor structure. The within-industry adoption path (industrial digitalization, proxied by e‑commerce sales) produces substantially larger effects than growth of the digital industries themselves (digital industrialization, proxied by telecommunications volume). The 2017 elevation of the digital economy to a national strategy marks a turning point: after 2017 digital industrialization’s net effect flipped (toward supporting industrialization and reducing servicization), while industrial digitalization continued to strengthen servicization and weaken industrial employment. Effects are stronger in urban areas than rural areas, and household-side digital adoption has a larger impact than government‑/enterprise‑side adoption.

Key Points

Overall effects (baseline, fixed‑effects models):
- Digital industrialization (telecommunications volume): industrialization β = −0.0084 (p<0.05); servicization β = +0.0099 (p<0.05).
- Industrial digitalization (e‑commerce sales): industrialization β = −0.0553 (p<0.01); servicization β = +0.0651 (p<0.01).
- Conclusion: both dimensions promote servicization and reduce secondary‑sector employment; industrial digitalization’s effect is larger.
Robustness:
- Alternative measures (share employed in IT services; computers per 100 persons) yield consistent signs and significance (large magnitudes reported).
Temporal heterogeneity (pre‑ vs post‑2017):
- Digital industrialization: pre‑2017 inhibited industrialization (β ≈ −0.0333) and promoted servicization (β ≈ +0.0324); post‑2017 it reversed (industrialization β ≈ +0.0182; servicization β ≈ −0.0202).
- Industrial digitalization: inhibited industrialization both periods (pre −0.0245; post −0.0392) and promoted servicization both periods (pre +0.0262; post +0.0432), with stronger effects after 2017.
Urban–rural heterogeneity:
- Urban digital economy: industrialization β = −0.0955; servicization β = +0.1105.
- Rural digital economy: industrialization β = −0.0352; servicization β = +0.0392.
- Urban impacts substantially larger.
Scenario heterogeneity (household vs government‑enterprise adoption):
- Household broadband: industrialization β = −0.0840; servicization β = +0.0984.
- Government/enterprise broadband: industrialization β = −0.0741; servicization β = +0.0828.
- Household‑side adoption shows stronger effects.

Data & Methods

Sample: provincial panel data for 31 provinces/regions in China, annual observations 2013–2024 (N = 372).
Data sources: National Bureau of Statistics of China, EPS Global Statistical Data Analysis Platform, CEInet, China Labor Economics Database, provincial statistical yearbooks.
Dependent variables (labor structure):
- Industrialization of labor structure = employment in secondary industry / total employment.
- Servicization of labor structure = employment in tertiary industry / total employment.
Key independent variables (two dimensions of digital economy):
- Digital industrialization (digital industries growth): total telecommunications business volume (primary measure); alternative: share of employees in information transmission/software/IT.
- Industrial digitalization (digital transformation within traditional industries): e‑commerce sales (primary measure); alternative: computers per 100 persons.
Heterogeneity measures: urban broadband users vs rural broadband users; household broadband users vs government‑enterprise broadband users; temporal split at 2017 (policy elevation of the digital economy).
Model: region and year fixed-effects panel regressions. Continuous variables (except ratios/dummies) log‑transformed. Coefficients interpreted as average percentage‑point changes in labor‑structure shares associated with digital indicators.
Robustness checks: alternative variable definitions; results reported as qualitatively consistent.

Implications for AI Economics

Path dependence matters: industrial digitalization (digital adoption within non‑digital sectors, including AI, e‑commerce platforms, automation) has a larger reallocation effect on labor than growth of pure digital sectors. For AI economics, this highlights that diffusion of AI across traditional industries may be the primary driver of employment mix shifts.
Servicization and deindustrialization tradeoffs:
- AI and broader digital adoption are associated with labor shifts toward services and away from secondary‑sector employment. Policymakers should plan for sectoral reallocation costs (retraining, mobility support).
- The stronger post‑2017 intensification of industrial digitalization effects suggests that policy and strategy (national promotion) can accelerate diffusion and amplify labor impacts.
Urban–rural and household vs firm effects:
- Larger urban impacts and stronger household‑side effects imply distributional consequences: urban workers and households benefit earlier/more from digital adoption, potentially widening urban–rural and within‑region inequalities. AI policies should prioritize rural connectivity, targeted skill programs, and inclusive adoption to avoid amplifying disparities.
Policy levers:
- Workforce development: focus on upskilling for service‑oriented and digitally intensive tasks, and on reskilling displaced industrial workers for higher‑value service occupations.
- Complementary investments: digital infrastructure (especially in rural areas), targeted subsidies for digital adoption in traditional industries that create local employment, social safety nets to smooth transitions.
- Monitoring and measurement: finer AI‑specific metrics (AI investment, firm‑level AI adoption, task automation indices) are needed to predict labor‑market impacts more precisely.
Research implications for AI economics:
- Need micro‑level causal evidence distinguishing task substitution vs task creation by AI across sectors and regions.
- Investigate effects on wages, job quality, hours, and inequality (not only employment shares).
- Explore dynamic adjustment: firm exit/entry, sectoral productivity gains, and long‑run labor demand elasticities as AI diffuses.
- Consider instrumenting for digital adoption or exploiting policy experiments to better identify causal pathways (e.g., infrastructure rollouts, targeted digital subsidies).

Short summary: the paper provides province‑level evidence that digitalization—especially the diffusion of digital technologies into traditional industries—has been an important driver of China’s shift toward a service‑oriented labor structure, with pronounced urban and household‑side heterogeneity and a marked change in patterns following the 2017 national strategy push.

Assessment

Paper Typequasi_experimental Evidence Strengthmedium — Uses long panel data and exploits a clear policy timing (2017) to analyze changes, giving credible temporal variation; however, causal interpretation is limited by potential reverse causality (regions shifting to services may adopt digital technologies), omitted time-varying confounders, and possible measurement error in digital economy indicators unless robust IVs or natural experiments are provided. Methods Rigormedium — Approach appears to use standard and appropriate panel methods and heterogeneity checks (urban/rural, household/government/enterprise applications) and examines a policy inflection point, but the description does not report quasi-experimental checks (parallel trends/event-study plots, placebo tests), instrumental variables, or micro-level validation that would raise rigor to high. SampleChinese provincial-level panel data covering years 2013–2024, with province-year observations measuring overall digital economy development and its two dimensions (digital industrialization and industrial digitalization); additional breakdowns by urban vs rural digital economy and by application side (household, government, enterprise). Themeslabor_markets adoption IdentificationProvince-level panel regressions with province and year fixed effects using variation in measures of digital industrialization and industrial digitalization over 2013–2024; the 2017 elevation of the 'digital economy' to a national strategy is treated as a policy shock (pre/post comparison and implied event-study/DID-style analysis). Identification therefore relies on temporal and cross-sectional variation and controls rather than randomized assignment or a clearly exogenous instrument. GeneralizabilityFindings are specific to China's institutional, policy and industrial context and may not generalize to other countries., Province-level aggregation may mask firm- and worker-level heterogeneity (ecological inference limits)., Results are conditional on the 2013–2024 period and the 2017 policy shock; different timing or technological trajectories elsewhere could yield different effects., Measures of the digital economy (indexes of digital industrialization vs industrial digitalization) may be context-specific and hard to replicate.

Claims (7)

Claim	Direction	Confidence	Outcome	Details
The development of the digital economy generally promotes the servicization and deindustrialization of the labor structure. Labor Share	mixed	high	servicization and deindustrialization of the labor structure (service and industry employment shares)	0.48
The driving effect of industrial digitalization on changes in the labor structure is stronger than that of digital industrialization. Labor Share	positive	high	relative magnitude of impact of industrial digitalization versus digital industrialization on servicization and deindustrialization	0.48
The elevation of the 'digital economy' to a national strategy in 2017 constituted a critical turning point in the relationship between digital-economy development and labor-structure change. Labor Share	mixed	high	change in the effect of digital-economy components on servicization and industrial employment shares after 2017	0.48
After 2017, digital industrialization shifted toward promoting industrialization and restraining servicization. Labor Share	mixed	high	post-2017 effect of digital industrialization on industrial employment share (industrialization) and service employment share (servicization)	0.48
After 2017, industrial digitalization continued to strengthen servicization while suppressing industrialization. Labor Share	mixed	high	post-2017 effect of industrial digitalization on service and industry employment shares	0.48
The urban digital economy exerts a stronger effect than the rural digital economy in promoting servicization and inhibiting industrialization. Labor Share	mixed	high	differential effect size of urban versus rural digital-economy development on service and industry employment shares	0.48
The impact of household-side digital economy applications on labor-structure change is significantly greater than that of government- and enterprise-side applications. Labor Share	positive	high	relative impact magnitudes of household- vs government- vs enterprise-side digital applications on service and industry employment shares	0.48