📈 economics

Zero-waste City Policy and the Short-Lived Efficiency Window in Urban Green Innovation: Evidence from China’s Staggered Pilot

Using a doubly robust difference-in-differences approach on Chinese prefecture-level data, this study finds that while China's Zero-waste City Pilot yields lasting improvements in industrial waste utilization, its positive impact on urban green innovation efficiency is short-lived, peaking only in the policy launch year and driven primarily by industrial restructuring.

Original authors: Peiwen Lu, Shouting Zhang, Kailingli Liao, Jinping Tian, Lyujun Chen

Published 2026-08-26

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Peiwen Lu, Shouting Zhang, Kailingli Liao, Jinping Tian, Lyujun Chen

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Cities are complex machines that consume resources and produce waste, but they are also engines of invention. For decades, environmental policy has focused on the mechanical side of this equation: how to collect trash, how to treat sewage, and how to keep the air clean. These are essential tasks, but they often treat waste as a problem to be contained at the very end of the process. A newer perspective suggests that how a city manages its waste can actually reshape how it invents the future. This idea sits at the intersection of environmental regulation and economic innovation. The core question is not just whether a city can clean up its act, but whether the effort to do so makes the city smarter and more efficient at turning ideas into useful, green technologies. If a city forces its factories to recycle more, does that pressure spark a wave of better inventions, or does it simply become a box-checking exercise that fades once the initial push is over?

In late 2018, China launched a massive experiment to test this very question. The government announced a "Zero-waste City" pilot program, a comprehensive plan designed to reduce waste at the source, recover resources, and upgrade industrial processes. The program did not roll out all at once; instead, it was introduced in waves, starting with eleven cities in 2019 and expanding to many more in 2022. This staggered timing created a natural laboratory for researchers to observe what happens when a city suddenly adopts these strict rules. A team of economists and geographers from China University of Geosciences and Tsinghua University set out to measure the true impact of this policy. They were not just counting how much trash was recycled; they were measuring the efficiency of the city's entire innovation system. They wanted to know if the policy made the city better at turning money and effort into green knowledge and economic value, and whether that improvement lasted.

To get a clear picture, the researchers looked at data from 210 cities across China spanning from 2012 to 2023. They built a detailed model that splits the innovation process into two distinct stages. The first stage is the creation of knowledge: how well a city turns research funding and scientists into new green ideas, measured by patent applications. The second stage is the conversion of those ideas into reality: how well the city turns those patents and energy into economic output while keeping pollution low. By tracking these two stages separately, the team could see exactly where the policy was working and where it was failing. They compared the cities that joined the pilot program against those that did not, using advanced statistical methods to account for the fact that the cities were different to begin with and that the policy arrived at different times.

The results revealed a surprising and fleeting pattern. When the first batch of cities launched the Zero-waste policy in 2020, there was an immediate and measurable jump in their innovation efficiency. In that single launch year, the cities became about 3.35% more efficient at turning inputs into green outputs. This boost was real and significant, affecting both the creation of new ideas and the conversion of those ideas into economic value. However, the story changed the moment the launch year ended. By 2021, 2022, and 2023, that initial boost had completely disappeared. The cities did not continue to get better year after year; instead, their efficiency levels drifted back down to where they were before the policy started. The overall effect of the policy, when averaged across all the years after the launch, was statistically zero. The policy had opened a short window of opportunity, but it did not build a lasting foundation for improvement.

This short-lived success stands in sharp contrast to the policy's performance on its most direct goal. While the efficiency of innovation faded, the actual rate of industrial waste recycling did not. The percentage of industrial solid waste that cities successfully reused and recycled jumped by nearly 6 percentage points in the launch year and continued to climb steadily in the following years. This suggests that the policy was highly effective at achieving its specific, administrative targets. The cities learned how to meet the rules, and they kept meeting them. But the ability to recycle waste did not translate into a permanent upgrade of the city's ability to innovate. The researchers found that the policy did not work by simply making companies pay more attention to environmental rules or by increasing the number of fines issued. Instead, the temporary boost in innovation efficiency came from a specific, structural shift: the policy caused a temporary reallocation of economic activity. It pushed cities to focus more on manufacturing and upgrading their industrial base relative to their service sectors during the launch period. This shift accounted for about 17% of the efficiency gain, but once that initial restructuring settled, the extra efficiency vanished.

The study concludes that while the Zero-waste policy successfully forced cities to clean up their industrial processes, it did not automatically create a self-sustaining engine for green innovation. The initial surge in efficiency was driven by intense administrative mobilization and a sudden reordering of resources, but without sustained market incentives and continued technology diffusion, that momentum could not be maintained. The cities proved they could comply with the rules, but they did not permanently become better at innovating. For policymakers, the lesson is clear: achieving a direct environmental target, like recycling more waste, is different from building a durable system of innovation. To make green innovation stick, cities need more than just a launch-year push; they need long-term support that helps them turn temporary coordination into permanent capacity.

Technical Summary: Zero-waste City Policy and the Short-Lived Efficiency Window in Urban Green Innovation

Problem Statement
While China's Zero-waste City Pilot (ZWCP) is designed to improve urban environmental performance, it remains unclear whether direct waste-governance achievements translate into durable improvements in urban green innovation efficiency (GIE). Existing literature often focuses on green patent counts or emission reductions, failing to distinguish between the quantity of innovation and the efficiency with which cities convert innovation inputs into green knowledge and environmentally adjusted economic output. Furthermore, few studies examine whether policy effects accumulate over time or concentrate in the initial implementation window, particularly within the context of staggered policy adoption where treatment effects vary across cohorts and time.

Methodology
The study utilizes a balanced panel of 210 Chinese prefecture-level cities from 2012 to 2023, covering 11 first-batch pilot cities (treated in 2020), 70 cities added in the 2022 expansion, and 129 never-treated control cities.

Efficiency Measurement (Two-Stage SBM-DEA):
The authors construct a composite GIE index using a two-stage slacks-based Data Envelopment Analysis (SBM-DEA) model under variable returns to scale (VRS).
- Stage 1 (Knowledge Production, $E_1$ ): Inputs include science expenditure and R&D employment; the desirable output is the total number of green patent applications.
- Stage 2 (Economic Conversion, $E_2$ ): Inputs include total energy consumption and green patent applications; the desirable output is GDP; undesirable outputs include industrial solid waste generation and industrial $SO_2$ emissions.
- The composite index is calculated as $GIE = 0.5E_1 + 0.5E_2$ .
Empirical Strategy (Staggered DID):
Recognizing that conventional two-way fixed effects (TWFE) estimators can be biased in staggered settings with heterogeneous treatment effects, the study employs the Callaway and Sant'Anna (2021) doubly robust difference-in-differences (CSDID) estimator with inverse probability weighting (DRIPW). This approach recovers group-time average treatment effects (ATT) while addressing the "not-yet-treated" control group contamination.
- Mechanism Analysis: A standard three-equation TWFE framework is used on a "clean" sample (11 first-batch pilots and 129 never-treated cities) to test mediation channels, as full mediation decomposition is not yet standard in multi-period CSDID settings.

Key Results

Short-Lived Efficiency Window: The ZWCP significantly increased GIE only in the launch year (2020) for the first-batch cities. The calendar-time ATT for 2020 is 0.0330 ( $p=0.043$ ), representing a relative gain of approximately 3.35%. However, effects in 2021–2023 are insignificant, and the overall average ATT across the post-treatment period is statistically indistinguishable from zero.
Stage Decomposition: The launch-year gain is observed in both stages of the innovation system. Knowledge-production efficiency ( $E_1$ ) increased by 4.64%, and economic-conversion efficiency ( $E_2$ ) increased by 2.26%. The gains in both stages fade after 2020.
Divergent Trajectories of Targets vs. Efficiency: While GIE showed a brief spike, the Industrial Solid Waste Comprehensive Utilization Rate (ISWCU) increased by 5.93 percentage points and continued to rise persistently through 2023. This divergence indicates that ISWCU is a successful direct policy target but does not act as a causal mediator for the short-lived GIE gains.
Mechanism Identification: Among the tested channels (ISWCU, government attention, enforcement intensity, and industrial restructuring), only industrial restructuring shows a statistically detectable indirect effect. The policy reduced the log ratio of tertiary-to-secondary value added (shifting resources toward the secondary sector), which accounted for approximately 17% of the total GIE effect. The direct effect of the policy on GIE, controlling for this channel, was insignificant.
Robustness: Results hold across seven robustness checks, including random permutation tests, announcement effect placebo tests, leave-one-out analyses, and controls for concurrent policies (e.g., Low-Carbon City pilots).

Significance and Contributions
The paper claims four primary contributions to the literature on environmental policy and urban sustainability:

Identification of a Launch-Year Efficiency Window: It challenges the assumption that green innovation effects persist automatically, demonstrating that ZWCP creates a temporary efficiency gain concentrated in the implementation year rather than a durable long-term trajectory.
Distinction Between Compliance and System Efficiency: The study differentiates between direct environmental targets (which can be sustained through administrative pressure, as seen with ISWCU) and broader innovation-system efficiency (which requires sustained technology diffusion and market incentives).
Methodological Decomposition: By applying a two-stage DEA framework within a staggered DID design, the paper isolates where efficiency gains arise (knowledge production vs. economic conversion) and identifies the specific timing of policy impacts.
Policy Evaluation Framework: The findings suggest that evaluating zero-waste policies solely on waste-utilization compliance is insufficient. Effective evaluation must track whether temporary administrative coordination is converted into stable innovation capacity.

Policy Implications
The authors argue that while the ZWCP successfully achieved direct waste-governance targets, its failure to sustain GIE improvements suggests that administrative mobilization alone is insufficient for long-term innovation efficiency. Future policy design should move from the "launch stage" (focused on targets and coordination) to "consolidation" and "diffusion" stages that prioritize R&D investment, technology adoption, and market incentives. Furthermore, because industrial restructuring was the primary channel, policymakers should manage the temporary shift toward the secondary sector through skills training and investment in cleaner production to ensure these structural changes translate into durable efficiency gains.

Limitations
The authors note that the small sample size of the first-batch pilots (11 cities) limits the precision of heterogeneity analysis, making those results descriptive rather than fully identified. Additionally, the 2020 launch coincided with the onset of the COVID-19 pandemic, which may have influenced the 2020 efficiency estimates, although interaction controls were used to mitigate this. Finally, the study is city-level; firm-level mechanisms remain to be explored.

More like this