0 cumulative citations
View corpus contextManufacturers that score higher on China's 'new quality productivity' index exhibit stronger total factor productivity, with gains partly driven by patents and digital transformation; the positive effect is robust across most regions but muted in the Northeast.
A study on the impact of new quality productivity on total factor productivity
openalex
quasi_experimental
medium evidence
7/10 relevance
Full text
usable extracted full text
DOI
Source
PDF
Using a panel of Chinese listed manufacturing firms (2015–2024), the authors find that higher firm-level 'new quality productivity' raises total factor productivity, with part of the effect transmitted through technological innovation (patents) and digital transformation (text-based index), and with heterogeneous regional effects.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
The transformation and upgrading of manufacturing have far-reaching effects on society and the economy. Manufacturing serves as the main engine of the national economy and the lifeblood of the real economy. Fully boosting the total factor productivity (TFP) of manufacturing is a key way to build a modern industrial system and encourage high-quality growth. Crucially, within the paradigm of sustainable industrial development, boosting manufacturing TFP through new quality productivity serves as a foundational micro-level engine for the transition toward a circular economy (CE), decoupling industrial growth from excessive resource consumption and environmental degradation. This study employs data from Chinese A-share listed manufacturing firms spanning the years 2015 to 2024 as its sample. Using a two-way fixed effects model, it looks at the effects of new quality productivity on corporate TFP and how these effects work. The results show the following: First, new quality productivity greatly increases manufacturing companies’ TFP. This conclusion holds even after several stability tests. Second, new quality productivity raises corporate TFP in two ways: through technological innovation and digital transformation. Third, heterogeneity analysis shows that new quality productivity exerts a statistically comparable and robust positive effect on firm TFP across the Eastern, Central, and Western regions, while its impact on the Northeastern region remains statistically insignificant, partly due to localized industrial structural rigidity and a smaller sample size.
Summary
Main Finding
New quality productivity (NQP) significantly increases firm-level total factor productivity (TFP) in Chinese manufacturing. This effect is robust to stability checks and operates partly through two mediating channels—technological innovation and digital transformation. Regional heterogeneity exists: positive effects are robust in Eastern, Central, and Western regions but statistically insignificant in the Northeastern region.
Key Points
- Definition and framing
- NQP is framed as a qualitative upgrade in productive forces (novelty, quality, force), driven by advanced technologies, higher-quality inputs, and new business/organizational models.
- TFP is treated as the firm-level efficiency metric for converting inputs into output; NQP is a directional, structural driver whose success is evaluated via changes in TFP.
- Main empirical results
- Baseline estimation (two-way fixed effects) finds a large, positive association between firm NQP and TFP.
- Mediation analysis shows:
- Technological innovation (proxied by log patent counts + 1) transmits part of NQP’s effect on TFP.
- Digital transformation (text-based index from frequency of digital keywords in annual reports, log+1) also mediates the NQP → TFP link.
- Heterogeneity: positive and comparable effects across Eastern, Central, Western China; Northeastern region effect is not significant (attributed to structural rigidity and smaller sample size).
- Identification and robustness
- Endogeneity addressed via multi-instrument 2SLS using three instruments: historical telecommunications endowment (1984 city-level postal/telecom volume × time trend), Confucian cultural heritage (count of historical Confucian academies × time trend), and city river density × time trend.
- Multiple control variables included (fixed asset ratio, growth capability, operating cash flow ratio, Tobin’s Q, capital intensity, ownership indicators, firm age, board composition, top-10 ownership share, etc.).
- Robustness tests and stability checks performed (details in paper).
Data & Methods
- Sample: Chinese A-share listed manufacturing firms, 2015–2024.
- Dependent variable: Firm-level TFP (used as the outcome measuring productivity; specific TFP estimation method not shown in the excerpt).
- Core explanatory variable: Firm-level New Quality Productivity (NQP). Constructed from a multi-indicator evaluation system (Table 1) covering dimensions such as workforce quality (R&D personnel salary share, share of highly educated personnel, R&D personnel proportion), labor object (fixed assets ratio, manufacturing overhead ratio), hard technology (R&D depreciation/amortisation ratios), etc.
- Mediators:
- Technological innovation: ln(total patent applications + 1).
- Digital transformation: textual frequency index of digital-related keywords from annual reports (ln(freq + 1)).
- Empirical strategy:
- Two-way fixed effects regression (firm and year/industry fixed effects).
- 2SLS with three historical/geographic instruments to mitigate reverse causality and omitted-variable bias.
- Controls as listed above; debt-to-equity ratio used as a potential moderator in additional analyses.
- Additional analyses: mediation tests for the two channels and regional heterogeneity checks.
Implications for AI Economics
- Mechanisms align with AI-driven productivity channels
- The mediating role of digital transformation and technological innovation is consistent with how AI (including machine learning and large models) can raise firm productivity: improved decision-making, automation, R&D acceleration, and product/process customization.
- Policy implications
- Policies that foster digital infrastructure, AI adoption, and human capital upgrading can magnify NQP and thus TFP gains—support for data infrastructure, AI-capacity building, and targeted R&D incentives is warranted.
- Regional tailoring: policies for lagging regions (e.g., Northeast China) should address structural industrial rigidity and build absorptive capacity for AI/digital technologies.
- Firm strategy
- Firms seeking productivity gains should invest in AI-enabled digital transformation and capability to commercialize technological innovations (patents, process redesign).
- Textual measures of digitalization (NLP of disclosures) are practical signals for monitoring digital adoption and can inform investment decisions.
- Methodological takeaways for AI economics research
- Combining text-based measures of digital adoption with conventional innovation metrics is a promising approach to quantify AI/digital impacts at firm level.
- Historical/geographic instruments (telecom endowment, cultural/human-capital proxies, geography) can help address endogeneity in studies of AI/digital diffusion—though external validity should be checked.
- Research opportunities
- Disaggregate which AI subcomponents (e.g., predictive analytics, generative models, automated control systems) drive the strongest TFP gains.
- Explore environmental/circular-economy outcomes of AI-driven NQP (the paper links NQP to circularity but micro-evidence could be deepened).
- Extend analysis beyond listed firms and outside China to assess generalizability and heterogeneous industry effects.
- Caveats relevant to AI economists
- The study focuses on listed manufacturing firms in China; effects may differ for SMEs, services, or other institutional contexts.
- Measurement choices (TFP method, NQP index construction, text-keyword lists) matter—sensitivity checks and replication with alternative measures are important.
- Instruments rely on historical and geographic persistence; researcher attention to instrument validity in other contexts is required.
If you want, I can: - Extract a concise table of the main variables and their operational measures from the paper; - Draft potential empirical strategies to isolate specific AI (large-model) impacts within the NQP framework; - Suggest keyword lists and NLP methods used for constructing the digital-transformation index.
Assessment
Paper Typequasi_experimental
Evidence Strengthmedium — The paper uses a large firm-level panel (Chinese A‑share manufacturing firms, 2015–2024), fixed effects, 2SLS, multiple robustness checks, and mediation analysis which provide reasonable empirical support. However, the validity of the proposed instruments (historical telecom endowment, Confucian academies, river density interacted with time trends) is questionable because they may correlate with persistent local economic traits that directly affect firm productivity; measurement of the NQP index and the digitalization proxy is also subject to construct and measurement error. These limitations reduce confidence in a clean causal interpretation.
Methods Rigormedium — The empirical strategy is standard and appropriate for panel productivity work (firm & time fixed effects, clustered/robust SEs implied, mediation tests, heterogeneity checks). Use of multiple IVs and 2SLS is a strength. Weaknesses include potentially weak or invalid exclusion restrictions for the instruments, likely measurement error in the constructed NQP index and textual digitalization measure, and no information here on first-stage strength, overidentification/weak-IV diagnostics, or placebo tests—making causal claims less secure.
SamplePanel of Chinese A‑share listed manufacturing firms, annual observations from 2015 to 2024; firm-level TFP as outcome, constructed index of 'new quality productivity' (NQP) based on multiple firm indicators (R&D personnel salary share, share of highly educated staff, fixed assets ratio, etc.), mediators include patent counts and a text-derived digital transformation keyword frequency index; controls include financial and governance variables.
Themesproductivity innovation adoption human_ai_collab
IdentificationFirm-level two-way fixed-effects panel regression with multi-instrumental-variable 2SLS to address endogeneity. Instruments: (1) historical (1984) city-level postal and telecommunications volume interacted with a time trend; (2) number of historical Confucian academies in the prefecture-level city interacted with a time trend; (3) city river density interacted with a time trend. Mechanism tests use mediation analysis with firm patent counts (technological innovation) and a text-based digital transformation index; controls for firm characteristics and heterogeneity analyses by region are included.
GeneralizabilitySample limited to publicly listed Chinese manufacturing firms (A‑share) — results may not generalize to private/smaller firms or services sector., China-specific institutional, industrial, and policy context (2015–2024) may limit external validity to other countries or periods., NQP is constructed from context-specific indicators and may not map cleanly to 'AI adoption' in other settings., Instruments rely on historical/geographic variation that may not translate beyond China or different eras.
Claims (5)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| New quality productivity significantly increases the total factor productivity of Chinese A-share listed manufacturing firms. Firm Productivity | positive | Firm-level total factor productivity |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Technological innovation mediates the positive relationship between new quality productivity and manufacturing firms’ total factor productivity. Firm Productivity | positive | Firm-level total factor productivity through technological innovation |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Digital transformation mediates the positive relationship between new quality productivity and manufacturing firms’ total factor productivity. Firm Productivity | positive | Firm-level total factor productivity through digital transformation |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The positive effect of new quality productivity on firm total factor productivity is statistically comparable and robust across China’s Eastern, Central, and Western regions. Firm Productivity | positive | Firm-level total factor productivity by geographic region |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The effect of new quality productivity on firm total factor productivity is statistically insignificant in China’s Northeastern region. Firm Productivity | null_result | Firm-level total factor productivity in the Northeastern region |
Reading fidelity
high
Study strength
low
|
not reported
|