Green taxation, regional green development and innovation: Mechanisms of influence and policy optimization

The study is motivated by the global shift toward green, low-carbon, and circular development and the growing policy focus on green taxation as a tool to internalize environmental externalities. Despite its prominence, theoretical and empirical gaps remain regarding green taxation’s adverse effects on regional green development and innovation, particularly via resource allocation and potential crowding out. Building on Resource-Based View (RBV), innovation systems, and technological lock-in theories, the paper investigates whether and how green taxes promote regional green development and innovation, while possibly inhibiting overall regional innovation capacity through reallocation of scarce resources and path dependence. The research formulates four hypotheses: H1: Green taxation exerts a delayed promotional effect on regional green development. H2: Green taxation positively influences regional green innovation. H3: Green taxation inhibits regional comprehensive innovation capacity. H4: Green taxation, while advancing regional green development and reallocating corporate resources, impedes the enhancement of regional innovation capacity.

The paper reviews the evolution and definitions of green taxation from Pigovian roots (Marshall, Pigou) to contemporary narrow (pollutant-focused) and broad (economy-ecology-society) conceptions. Prior studies indicate green taxes can reduce emissions, encourage environmental technology adoption, and support sustainable development (e.g., Pourkarimi & Hojjat; Chien et al.; Hu & Chen; Labandeira et al.; Rodríguez et al.). Green innovation is defined as process and product innovations that mitigate environmental problems (Franceschini; Kemp & Oltra; Rennings), driven by regulation, stakeholder pressure, and cost efficiency. Evidence suggests environmental taxes and related policies can stimulate green innovation (Porter hypothesis; Wang et al.; Yu et al.; Jiang et al.; Li & Li; Zheng et al.; Deng et al.). However, the literature underexplores potential negative effects: resource reallocation away from other innovation domains (RBV: Zahra), systemic shifts altering incentives and flows (innovation systems: Kooreman & Mot), and lock-in constraints (Foxon; Cecere et al.). The review highlights the need to examine both positive and negative mechanisms of green taxation on regional green development and innovation, including nonlinear effects and regional heterogeneity.

Data: Unbalanced panel of 30 Chinese provinces/municipalities from 2004–2021. Green tax intensity and pollution fee data are from the China Environment Yearbook; other economic and control variables from China Statistical Yearbook and provincial yearbooks. Missing data are imputed by regression. All variables are log-transformed. Models: Two-way fixed effects (province and year) panel regressions with clustered robust standard errors. F-tests and Hausman tests support fixed effects over pooled or random effects. Stata16 is used. Explanatory variables: Green taxation split into narrow (LNXYGTAX: environmental protection/pollution fees share of total regional tax) and broad (LNGYGTAX: share of environmental protection tax, domestic consumption tax, resource tax, urban maintenance and construction tax, farmland occupation tax, and vehicle & vessel tax in total regional tax). Dynamic specifications include contemporaneous, one-period (L.) and two-period (L2.) lags and squared terms to capture potential nonlinearities. Dependent variables: LNGTTP (regional green development), measured by SBM-DDF green total factor productivity using inputs: energy consumption, fixed assets (perpetual inventory method), employment; desired output: GDP; undesired outputs: industrial SO2, wastewater, particulate matter. LNGTC (regional green innovation), measured as ratio of green patent grants to total patent grants (log-transformed). LNTCC (regional comprehensive innovation capacity), from China Regional Innovation Capability Report (2004–2021) aggregation across five primary dimensions: knowledge acquisition (0.15), knowledge creation (0.15), enterprise innovation (0.25), innovation environment (0.25), innovation performance (0.20). Control variables: education level (LNEDU: education expenditure share), openness (LNOP: trade/GDP with exchange rate adjustment), economic development (LNGDP: per capita GDP), technological innovation (LNTC: fiscal R&D expenditure share), foreign direct investment (LNFDL), labor (LNL: employed persons at year-end), industrial structure (LNIS: tertiary industry share of GDP). Empirical strategy: Baseline two-way FE regressions for narrow and broad taxes on LNGTTP, LNGTC, LNTCC with dynamic and quadratic terms. Mechanism (mediation) analysis follows Wen & Ye’s three-step approach, introducing LNGTTP as mediator in LNTCC equations to assess whether green development mediates tax effects on innovation capacity. Heterogeneity analyses: by region (Eastern, Central, Western), by period (before 2006; 2006–2016; after 2016), and by innovation capacity (high vs. low regional comprehensive innovation capacity groups). Robustness checks: alternative estimators (random effects, OLS), and alternative green tax measure (total amount vs. intensity) to address measurement concerns; results remain consistent.

• Baseline results (two-way FE): Narrow green taxes show a significant positive lagged effect on green development (first lag significant at 1%). Narrow taxes are positively associated with green innovation at 10% (contemporaneous), but the first lag exhibits a negative effect at 10%, indicating short-term stimulation with subsequent cost-dominant effects. Narrow taxes significantly reduce comprehensive regional innovation capacity with a one-period lag (5%). Broad green taxes have limited or insignificant effects on green development and green innovation at the national level, but significantly reduce comprehensive innovation capacity with both first and second lags at 10%; the quadratic term is significantly negative, implying increasing inhibition as broad taxes rise. Selected coefficients from Table 8 (t-stats in parentheses): L.LNXYGTAX on LNGTTP: 0.104 (2.85); LNXYGTAX on LNGTC: 0.161 (1.78) with L.LNXYGTAX −0.055 (−1.95)*; L.LNXYGTAX on LNTCC: −0.031 (−2.28); LNGYGTAX^2 on LNTCC: −0.041 (−1.72)*. R-squared around 0.63–0.77 across specifications. - Mechanism (mediation) tests: Introducing LNGTTP into LNTCC regressions weakens the inhibitory effect of narrow taxes on innovation capacity (from 5% to 10%), and LNGTTP itself negatively affects LNTCC at 5% (coefficient ≈ −0.079 to −0.081), indicating that green development can mediate and transmit a negative effect to comprehensive innovation capacity (supporting H4). - Correlation analyses: Narrow green taxation correlates negatively with green development, but the squared term is positive, suggesting a U-shaped relationship; broad green taxation correlates negatively with comprehensive innovation capacity, with a negative squared term, indicating an inverted U-shaped inhibitory pattern. - Regional heterogeneity (narrow taxes): Eastern region—narrow taxes significantly promote green development (5%), with minimal effect on comprehensive innovation capacity; they inhibit green innovation (10%). Central region—weak effect on green development; significant inhibition of green innovation (5%) and negative effects on enterprise innovation capacity (10%). Western region—weak effect on green development; promote green innovation with a first-order lagged negative effect (10%); significant negative impact on comprehensive innovation capacity (5%), with a significant negative quadratic term. - Regional heterogeneity (broad taxes): Eastern—first-order lagged positive effect on green development (1%) and positive effect on green innovation (10%); inhibit comprehensive innovation capacity with a significant negative quadratic term. Central—enhance green innovation (1%) but show inverted U-shaped effects; mixed effects on comprehensive innovation capacity (promotional and first-lag inhibitory) with an inverted U-shape; limited impact on green development. Western—broad taxes tend to impede green development and comprehensive innovation capacity (10%); effects on green innovation are generally insignificant. - Temporal heterogeneity: Narrow taxes had limited influence on green development before 2016 but became significantly promotional afterward (1%); their negative effect on innovation capacity persists across periods, with stronger quadratic inhibition post-2006. Broad taxes showed promotive effects on green development and inhibitory effects on green innovation before 2006 (both 1%); from 2006–2016, diversified environmental policies attenuated tax impacts on green development and strengthened the negative effect on innovation capacity; post-2016, negative effects on innovation capacity remain but gradually weaken amid supportive policies. - Innovation-capacity heterogeneity: Narrow taxes promote green development in both high- and low-innovation regions (lagged), inhibit green innovation in high-innovation regions (5%), and promote green innovation in low-innovation regions (1%); they hinder comprehensive innovation capacity mainly in low-innovation regions (5%). Broad taxes do not significantly promote green development or innovation in either group; they significantly inhibit comprehensive innovation capacity in high-innovation regions (first and second lags at 1%), while effects are not significant in low-innovation regions (partly offset by supportive policies). - Overall: H1 and H2 are supported to varying extents; H3 is strongly supported; H4 is supported by mediation results showing green development can transmit a negative effect to comprehensive innovation capacity via resource reallocation/crowding-out.

The findings reconcile the apparent paradox of green taxation by showing that while green taxes can be effective levers for regional green development and, under certain conditions, green innovation, they can simultaneously suppress comprehensive innovation capacity. The mechanism aligns with RBV and innovation systems perspectives: finite resources are reallocated toward compliance and green-specific R&D, potentially crowding out other innovative activities and reducing overall regional innovation capacity. Technological lock-in further suggests that firms and regions may double down on existing technological trajectories, limiting exploration of broader innovation avenues. Nonlinearities (U-shaped for narrow taxes with green outcomes; inverted U-shaped for broad taxes with innovation capacity) indicate that effects vary with tax intensity and design. Heterogeneity across regions (Eastern vs. Central vs. Western) and across innovation-capacity strata underscores that institutional readiness, industrial structure, and existing capabilities condition the efficacy and side effects of green taxes. The mediation analysis shows that even as green development improves, it may temporarily depress comprehensive innovation capacity through resource diversion and transitional industry restructuring. Policy coherence (e.g., coupling taxes with subsidies/reinvestment, differentiated regional designs) is therefore essential to amplify benefits and mitigate innovation-capacity losses.

The study contributes by: (1) distinguishing narrow and broad green taxes and quantifying their differential effects on regional green development, green innovation, and comprehensive innovation capacity using a two-way fixed effects framework with dynamic and nonlinear terms; (2) demonstrating that green taxes promote green development (notably with lags) and conditionally stimulate green innovation, but both narrow and broad taxes significantly suppress comprehensive innovation capacity at the regional enterprise level; (3) uncovering nonlinear relationships (U-shaped and inverted U-shaped) and pronounced heterogeneity across regions, time, and innovation-capacity levels; (4) identifying a mediating mechanism whereby green development can transmit a negative effect to comprehensive innovation capacity via resource reallocation. Policy implications: - Strengthen targeting in green tax design, with reliefs and incentives for firms proactively adopting clean energy, recycling, and low-carbon processes; optimize rates and administration to avoid excessive burdens. - Substantially bolster support to Central and Western regions through fiscal subsidies, special funds for green R&D and environmental management, and complementary policies to raise capacity and reduce compliance costs. - Reinvest a portion of green tax revenues into enterprise green R&D, environmental management, and resource recycling to offset negative impacts on comprehensive innovation capacity. Future research: Examine temporal dynamics and lags between R&D and commercialization; evaluate optimal tax rates and levy methods; analyze interactions among green taxes, subsidies, and market mechanisms (e.g., carbon markets) to identify policy mixes that maximize green innovation while preserving overall innovation capacity.