When a Measure Becomes a Target: China’s Environmental Tax Spurs Low-Quality Green Patents, Not Breakthroughs

When China converted its pollution discharge fees into a formal environmental tax in 2018, policymakers hoped the harder-edged fiscal instrument would push heavily polluting firms toward genuine clean-technology breakthroughs. A new study suggests something more subtle happened instead: the companies most exposed to the tax became visibly greener on paper, but the improvement was concentrated in the cheapest, fastest, and least demanding form of green patenting — and in accounting choices that made their research spending more conspicuous. The findings, published in Frontiers in Environmental Economics, offer a striking empirical demonstration of Goodhart’s Law in environmental policy: when a measure of innovation becomes a target, it may cease to measure innovation.

The study, conducted by Lifan Pang of the Business School at the University of Chinese Academy of Social Sciences, exploits China’s 2018 Environmental Fee-to-Tax Reform as a quasi-natural experiment. Before the reform, pollution discharge fees were collected by local environmental authorities under a system long criticized for weak enforcement, discretionary implementation, and local protectionism. Firms could sometimes bargain over payments or delay them. The Environmental Protection Tax Law, which took effect on January 1, 2018, replaced this arrangement with a statutory tax backed by standardized collection procedures and administered with the involvement of tax authorities. The change strengthened the legal basis of pollution charges, increased enforcement rigidity, and made environmental compliance far more visible to regulators, investors, and other stakeholders.

Because emission intensity varies enormously across industries, the reform did not hit all firms equally. Companies in heavily polluting sectors were far more likely to emit taxable pollutants — air pollutants, water pollutants, solid waste, and noise — and to face higher compliance costs. This asymmetry gave the researcher a clean comparison group: heavily polluting listed firms could be matched against other Chinese A-share listed firms before and after the policy shock. To make the two groups comparable, the study employs an entropy-balanced difference-in-differences design, constructing weights from firm-level pre-treatment characteristics measured between 2013 and 2017. After balancing, standardized biases between treated and control firms fall to between 0.001 and 0.003, reductions of roughly 96 to 99.5 percent, meaning the weighted control group closely mirrors the treated group on observable dimensions such as size, leverage, profitability, growth, cash flow, and governance.

The analysis draws on a panel of Chinese A-share listed firms from 2013 to 2024, combining financial and governance data from the China Stock Market & Accounting Research Database with green patent data from the Chinese Research Data Services Platform. The initial sample of 5,627 firms and 55,845 firm-year observations is narrowed, after excluding financial firms, distressed ST/PT firms, and records with missing core controls, to 4,580 firms and 46,825 observations. Firms lacking valid pre-reform covariates cannot receive entropy-balancing weights, leaving a final estimation sample of 2,866 firms and 31,090 firm-year observations — a reduction driven by data availability rather than by outcomes. Firm-years without matched green patent records are coded as zero applications rather than dropped, and all continuous variables are winsorized at the 1st and 99th percentiles.

The first major result concerns how firms account for research and development. After the reform, heavily polluting firms significantly increased R&D expensing — the recognition of research spending as a current-period cost — while reducing capitalized R&D, the treatment of development expenditure as a long-term asset on the balance sheet. The difference-in-differences coefficient for R&D expensing is 0.251, significant at the 1 percent level, while the coefficient for capitalized R&D is −0.622, also significant at the 1 percent level. The R&D capitalization ratio declines significantly, and the ratio of expensed to capitalized R&D rises sharply. In other words, the reform shifted the composition of R&D recognition toward more immediately visible expense reporting. The author is careful to frame this as strategic R&D recognition rather than direct evidence of fraudulent accounting manipulation: capitalization decisions involve managerial judgment about project uncertainty and expected future benefits, and a shift toward expensing may reflect both genuine conservatism and a desire to display innovation effort more prominently.

The second major result concerns green patenting, and here the quality distinction becomes central. Chinese patent law recognizes two relevant categories: invention patents, which require higher novelty and inventiveness, undergo substantive examination, and take longer to obtain; and utility model patents, which are easier, faster, and cheaper to secure. If environmental taxation were inducing substantive green innovation, one would expect a robust increase in invention patents. Instead, the reform increased total green patent applications, with a coefficient of 0.114, but the increase was concentrated in utility model patents, with a coefficient of 0.185, both significant at the 1 percent level. Green invention patents showed a coefficient of just 0.028, statistically insignificant. A Wald test decisively rejects the hypothesis that the two coefficients are equal, and the utility model share of green patents rose significantly while the invention share did not.

Because patent counts contain many zeros, the standard logarithmic transformation can distort estimates, so the study subjects its patent results to three alternative estimators: Poisson pseudo-maximum likelihood on raw counts, the inverse hyperbolic sine transformation, and extensive-margin linear probability models. Across all three, the pattern holds — utility model patents respond strongly and significantly while invention patents remain statistically insignificant, with p-values ranging from 0.161 to 0.332 for the invention outcome. The study also rules out the possibility that the R&D expensing result is an artifact of disclosure selection: a linear probability model shows the reform did not systematically change which firms disclose R&D expenses in their management discussion and analysis sections, and re-estimating all key outcomes on the same restricted sample preserves the asymmetric pattern. Placebo tests based on 500 random treatment assignments confirm that the actual estimates for expensing, capitalization, and utility model patenting are unlikely to arise by chance, while the invention patent estimate does not stand out from the placebo distribution.

Perhaps the most revealing findings come from the ownership heterogeneity analysis. State-owned enterprises, which face stronger political and legitimacy pressures and are expected to align with public policy priorities, showed the strongest increases in R&D expensing and green utility model patenting — a pattern the author interprets as symbolic compliance, improving visible indicators of environmental effort without a corresponding surge in substantive invention. The SOE total effect on R&D expensing is 0.396 and on utility model patents is 0.325, both significant at the 1 percent level, yet the SOE total effect on invention patents is statistically insignificant. Privately owned enterprises, which typically have weaker access to policy resources and external financing, responded differently: they mainly reduced capitalized R&D, with a coefficient of −0.915, consistent with defensive contraction under higher compliance-cost pressure, while showing no significant patent response at all. Supplementary within-firm analysis shows that residualized R&D expensing and residualized utility model patenting co-move positively, stronger among SOEs than among private firms, suggesting that accounting-side and patent-side signals may form part of a coordinated strategic compliance portfolio — though the author emphasizes this association is correlational, not causal.

The findings qualify, rather than refute, the Porter Hypothesis — the influential idea that well-designed environmental regulation can offset compliance costs by stimulating innovation and improving productivity. The reform did increase measured green patenting, but the marginal response was disproportionately concentrated in low-threshold patent types. Aggregate patent counts, the study suggests, may overstate the substantive innovation effect of environmental regulation when the increase is driven by utility models rather than inventions. The author also notes that substantive innovation and strategic patenting are not mutually exclusive; firms may engage in both, and the empirical question is whether the marginal response tilts toward the symbolic. The evidence indicates it does, though this does not imply that all utility model patents are meaningless or that all firms engage in greenwashing.

The policy implications are direct. Regulators, the study argues, should not rely solely on aggregate green patent counts or total R&D expenditure when evaluating firms’ environmental innovation performance. Evaluation systems should complement quantitative metrics with quality-sensitive indicators — invention patents, patent citations, patent maintenance, commercialization outcomes, and actual environmental performance. When measurable signals become targets, firms will optimize the signals; only by weighting quality alongside quantity can policymakers reduce the incentive for symbolic compliance and steer environmental taxation toward the substantive green innovation it was designed to provoke.

Subject of Research: How China's 2018 environmental tax reform affected R&D accounting choices and the quality of green patenting among heavily polluting listed firms

Article Title: Not all innovations are created equal: environmental taxes, strategic R&D accounting, and green patent quality in China

Article References: Pang, L. (2026). Not all innovations are created equal: environmental taxes, strategic R&D accounting, and green patent quality in China. Frontiers in Environmental Economics, 5, Article 1894951. https://doi.org/10.3389/frevc.2026.1894951

Image Credits: AI Generated

DOI: 10.3389/frevc.2026.1894951

Keywords: environmental taxation, green innovation, patent greenwashing, utility model patents, R&D capitalization, R&D expensing, difference-in-differences, entropy balancing, state-owned enterprises, China, Porter Hypothesis, environmental regulation