How does industry-university-research collaborative innovation affect energy intensity in China: a novel explanation based on political turnover

China, the largest carbon emitter, plays a critical role in global emissions reduction. In 2021 China’s CO₂ emissions reached 10,523 million tonnes (31.1% of global emissions). As part of its Paris Agreement commitments, China targets a 60–65% reduction in CO₂ emissions per unit of GDP from 2005 levels and peaking emissions around 2030. Energy intensity, defined as energy consumption per unit of output, is a key indicator of energy efficiency; reducing it can cut emissions while sustaining growth. Technological innovation can substantially reduce energy intensity, yet many firms in developing countries lack strong internal R&D capacity. IUR collaborative innovation integrates the complementary roles of industry (commercialization), universities, and research institutes (knowledge creation) to overcome innovation constraints. In China’s context, government influence and political dynamics are pivotal; political turnover may disrupt or reshape policies and resource allocation, potentially affecting how IUR collaboration translates into energy intensity outcomes. This study asks: (1) How does IUR collaborative innovation affect energy intensity? (2) Does political turnover moderate this relationship? Findings aim to inform policy for achieving China’s dual-carbon goals.

Prior work identifies multiple drivers of energy intensity, including economic growth, energy prices, trade openness, and policies. Technological innovation generally lowers energy intensity, though evidence on direct effects is mixed. Collaborative innovation—positioned between closed and open innovation—enhances innovation capacity under resource constraints and can improve efficiency and reduce transaction costs, potentially lowering energy intensity. However, most studies proxy IUR collaboration using outputs (patents, publications), which overlook inputs and do not capture the distinct contributions of industry, universities, and research institutes. Government factors also shape innovation, but research has focused more on policy instruments than on officials’ characteristics. Political turnover influences environmental policy implementation, firm behavior, and innovation incentives via promotion tournaments and anti-corruption effects, yet its role in the IUR–energy intensity link remains underexplored. The authors hypothesize: H1, IUR collaborative innovation reduces energy intensity; H2, political turnover positively moderates this effect.

Data: Panel of 30 provincial-level regions in mainland China (excluding Tibet, Hong Kong, Macao, Taiwan) from 2010–2018, covering over 90% of China’s population and GDP and including diverse regional types. Energy consumption and IUR indicators come from the China Energy Statistical Yearbook and China Science and Technology Statistical Yearbook; RD intensity, population density (PD), foreign direct investment (FDI), and industrial structure (IS) come from the China Statistical Yearbook. Political turnover data (provincial party secretaries) come from provincial government websites. Missing values in some years/provinces were supplemented from the National Bureau of Statistics and provincial reports. Monetary variables are deflated to 2010 constant prices. Variables: - Dependent: Energy intensity (EI) = primary energy consumption (PEC) / gross regional domestic product (GRDP). - Key independent: IUR collaborative innovation (CI), measured via a synergetic composite system model capturing inputs and outputs of three subsystems (university, industry, research institute). Indicators include R&D personnel FTE, internal R&D expenditure, R&D projects (industry), scientific papers, patent applications, and sales revenue of new products (industry). Indicator weights are determined by an entropy method (time-adjusted). Degrees of order are computed for each subsystem and synthesized to a synergy degree for IUR collaboration. - Moderator: Political turnover (PT), a dummy for newly appointed provincial party secretary. Following prior work, PT=1 if the secretary assumes office between Jan 1–June 30 of year t (allowing policy effects to manifest), else 0. - Controls: R&D expenditure input intensity (RD, R&D/GRDP), population density (PD, log), FDI (log net inflows), industrial structure (IS, share of secondary industry in GRDP). Models: - Baseline panel regression: EI_it = β0 + β1 CI_it + β2 RD_it + β3 PD_it + β4 FDI_it + β5 IS_it + ε_it. - Moderation model adds PT and interaction: EI_it = λ0 + λ1 CI_it + λ2 PT_it + λ3 (CI_it×PT_it) + controls + ε_it. Primary parameters of interest: β1 (effect of CI) and λ3 (moderation by PT). - Estimation strategies: OLS, Fixed Effects (FE), Random Effects (RE) with Hausman tests guiding FE vs RE; dynamic specification using System GMM (SYS-GMM) with lagged EI as an instrument to address endogeneity and persistence. Robustness: - Instrumental variable approach using the historical opening of commercial ports as an instrument for innovation openness/history (relevance) and plausibly exogenous timing (exogeneity). - Variable substitution (e.g., patent counts as alternative IUR proxy), grouped regressions by PT, counterfactual timing of PT (advancing by 1–2 years), and inclusion of policy dummies (carbon trading pilots, low-carbon pilots). Multiple checks (FE-2SLS, SYS-GMM with Hansen and AR(2) tests) support robustness. Sample descriptive stats (N=270): mean EI=0.830 (SD=0.436), CI mean synergy=0.005 (SD=0.006), PT=0.285; RD mean=1.583%; PD mean≈2840 persons/km²; FDI mean≈$8.02B; IS mean≈44.9%.

- IUR collaborative innovation reduces energy intensity: In baseline regressions, CI has negative coefficients—OLS: −0.0582 (t=−2.63, p<0.01); FE: −0.0217 (t=−1.91, p<0.10); RE: −0.0265 (t=−2.34); SYS-GMM: −0.0113 (t=−1.83). Supports H1. - Political turnover strengthens the EI-reducing effect of IUR collaboration: Interaction CI×PT is negative and significant—FE: −0.0131 (t=−1.87, p<0.10); SYS-GMM: −0.0111 (t=−1.98, p<0.05). Supports H2. - Regional heterogeneity: • Eastern region: CI negatively affects EI (FE: −0.0430, t=−1.71), indicating stronger IUR synergy and market conditions facilitate EI reduction. • Central and Western regions: Direct CI effect is not significant; however, PT significantly enhances CI’s negative effect on EI—Central FE: CI×PT = −0.0188 (t=−2.09, p<0.10); Western RE: CI×PT = −0.1459 (t=−2.07, p<0.05). - Evaluation of IUR synergy (2010–2018): Average IUR synergy degree is ~0.0055; higher in 2012 (0.0066) and 2018 (0.0067); lower in 2010 (0.0040) and 2015 (0.0049). University subsystem shows higher order degree (approx. 0.0769–0.1557) than industry and research subsystems. Top provinces by average synergy: Jiangsu (0.0212), Guangdong (0.0185), Beijing (0.0154). Many western and some central provinces remain below average. - Controls and diagnostics: RD often shows a negative association with EI in FE; FDI and IS display expected signs in several specifications. Hausman tests favor FE. Dynamic models pass Hansen (e.g., p≈0.15–0.36) and AR(2) tests, supporting instrument validity and model specification. Robustness checks (IV, variable substitution with patents, grouped regressions, counterfactual PT timing, policy controls) consistently corroborate main results.

The findings directly address the research questions: IUR collaborative innovation, by integrating knowledge creation (universities, research institutes) with commercialization (industry), enhances technological progress and reduces transaction costs, leading to lower energy consumption per unit of output, thereby reducing energy intensity. Political turnover further strengthens this effect, consistent with China’s promotion tournament incentives wherein newly appointed officials prioritize energy-saving performance and allocate resources to innovation, while reduced collusion during turnover periods can improve resource efficiency. Regional disparities reflect different marketization levels, innovation ecosystems, and energy resource endowments: the eastern region’s mature markets and stronger IUR networks yield more immediate EI reductions; in central and western regions, political turnover is a key catalyst that enhances the effectiveness of IUR collaboration under tighter resource and growth pressures. These results underscore the importance of both collaborative innovation ecosystems and governance dynamics in achieving energy efficiency and progressing toward dual-carbon goals.

This study constructs an input–output-based synergy index for IUR collaborative innovation and empirically examines its impact on energy intensity across 30 Chinese provinces (2010–2018), while introducing political turnover as a moderating factor. Main contributions: (1) IUR collaborative innovation significantly reduces energy intensity; (2) political turnover amplifies this effect, especially outside the highly marketized eastern region; (3) notable regional heterogeneity highlights the eastern region’s stronger direct benefits from IUR collaboration and the central/western regions’ dependence on political dynamics to realize gains. Policy implications: prioritize construction of IUR collaboration platforms and bases (especially in central/western regions), strengthen resource-sharing and talent mobility mechanisms, and support R&D capabilities; newly appointed provincial leaders should design policies that incentivize and facilitate IUR collaboration for energy efficiency; refine officials’ evaluation systems to incorporate sustained energy efficiency and innovation outcomes. Future research may disaggregate political turnover types (promotion/leveling/demotion) and extend analysis to municipal-level data to capture finer spatial dynamics.

- Political turnover is modeled as a binary indicator for newly appointed provincial party secretaries within a fixed window; it does not distinguish between promotion, lateral moves, or demotion, which may have different incentive effects. - The analysis is at the provincial level; municipal-level studies could reveal more granular mechanisms and spatial heterogeneity. - Some provincial-year data were missing in yearbooks and were supplemented from alternative official sources; while authoritative, such imputation may introduce measurement uncertainty. - The IUR synergy index relies on available indicators and entropy-weighting; although it captures both inputs and outputs across subsystems, unobserved dimensions of collaboration quality may remain.