A policy designed to scrub livestock farms out of environmentally sensitive zones in southern China has achieved its local goal, but at a cost that ripples far beyond the boundaries where it was enforced. A new study of Guangdong Province, published in Carbon Research, finds that the so-called Livestock-Free Zone policy reduced carbon emissions inside regulated areas while simultaneously pushing an even larger share of emissions into neighboring counties. According to the spatial econometric analysis, the spillover effect was approximately 3.14 times larger than the local reduction, a disparity the authors describe as policy-induced leakage.
The research team, led by Zhiyi Ye and Xunan Yang, analyzed county-level panel data covering 2014 through 2023, combining a spatial Durbin model, a continuous difference-in-differences design, and network analysis. Livestock and poultry farming is a major source of methane and nitrous oxide, two greenhouse gases with far higher warming potential per unit than carbon dioxide, and the sector has often been overshadowed in China’s decarbonization agenda by industrial and energy policy. The Livestock-Free Zone approach, adopted widely by local governments, prohibits breeding within designated areas and functions, in the authors’ words, as a form of regional shock therapy that rapidly clears targeted territories.
The carbon accounting followed the IPCC Guidelines for National Greenhouse Gas Inventories, using the 2006 edition with its 2019 revision, and drew on the EDGAR v8.0 database and China’s national greenhouse gas inventory. The scope covered methane from enteric fermentation and both methane and nitrous oxide from manure management, calculated from annual slaughter and inventory numbers for pigs, cattle, sheep, and poultry in each county. Policy intensity was constructed as a time-varying continuous variable: the ratio of exclusion-zone area to total administrative area, updated whenever local governments adjusted zone boundaries. Where official area figures were unavailable, the team extracted exclusion-zone pixels from public image datasets using Python’s OpenCV library or performed spatial overlay analysis in ArcGIS, with validation against government measurements showing a relative error of about 4.19 percent.
The headline numbers come from decomposing the spatial Durbin model into direct and indirect effects. The direct effect of the policy on local livestock carbon emissions was negative and significant, at minus 0.047, confirming that stricter exclusion measures did cut emissions where they were enforced. The indirect, or spillover, effect was positive and significant, at plus 0.148, meaning that a 1 percent increase in policy intensity in one region was associated on average with a 0.148 percent increase in livestock carbon emissions in neighboring regions. Because the absolute value of the spillover dwarfs the local reduction, the study concludes that the environmental burden displaced by the policy far exceeds the emissions it eliminated, a pattern the authors link to the classic Pollution Haven Hypothesis, in which polluting industries migrate from strictly regulated jurisdictions to lenient ones.
A continuous difference-in-differences model, which accommodates variation in policy strictness across regions rather than treating the policy as a simple on-off treatment, reinforced the causal picture. The core policy variable carried a significantly negative coefficient of minus 0.035. An event-study framework following Sun and Abraham confirmed parallel trends before the policy shock and revealed a distinct temporal pattern: the local reduction effect peaked within the first two years of enforcement and then gradually dissipated by the third year, consistent with high-emission production capacity relocating quickly to loosely regulated neighbors during the initial phase.
The geography of the transfer is strikingly hierarchical. Total livestock emissions in Guangdong fell from 11.809 to 9.081 million tonnes over the study period, and emissions intensity dropped from 3.51 to 1.35 million tonnes per unit of livestock GDP, yet the spatial pattern restructured dramatically. Emissions in the Pearl River Delta, the province’s most developed and most strictly regulated region, declined steadily from 3.149 to 2.017 million tonnes, turning it into a low-emission zone. Northern Guangdong, by contrast, achieved only a modest total reduction of 0.347 million tonnes, and the cities of Shaoguan and Qingyuan actually saw emissions rise by 0.133 and 0.067 million tonnes respectively between 2018 and 2023, the only increases in the province.
Network analysis built from the model’s spillover matrix identified the sources and sinks of this flow with precision. The strongest input effects, measured as weighted in-degree, concentrated in northern Guangdong, with Shaoguan at 27.6 and Qingyuan at 27.3 serving as the province’s primary receptors. Jiangmen and Guangzhou in the Pearl River Delta emerged as the principal output sources, while Zhanjiang, Zhaoqing, and Jieyang formed secondary input centers. Maoming in western Guangdong proved a distinctive case, functioning simultaneously as a major input node and a secondary transmitter of spillovers into surrounding areas. A sample perturbation test confirmed that removing the Pearl River Delta from the analysis flipped both direct and indirect effects negative, establishing that the delta’s policy shock was the primary driver of the transfer.
The consequences for recipient regions were not uniform, and this asymmetry forms one of the study’s most important findings. Western Guangdong, anchored by Zhanjiang and Maoming, absorbed new capacity while maintaining high production efficiency, displaying a pattern of high total emissions but low carbon-emission intensity thanks to long-established large-scale, intensive farming. Northern Guangdong, the major spillover recipient, fell into what the authors call an efficiency dilemma: capacity absorption without sufficient efficiency gains. According to 2025 statistics from the Guangdong Provincial Department of Agriculture and Rural Affairs cited in the study, northern Guangdong has only half as many leading enterprises as western Guangdong, at 38, and merely one-third of the agricultural patents, at 430, yet it accounts for four of the five cities in Guangdong with the highest carbon emission intensities. Mountainous terrain hampers the construction of large standardized facilities there, so much of the absorbed capacity settled into dispersed, small-to-medium operations without synchronized upgrades to farming practice or environmental protection technology.
The authors identify two mechanisms behind the transfer. The first is a market substitution effect: when policy restrictions shrink local livestock output in core consumption markets such as the Pearl River Delta, unmet demand pushes neighboring regions with weaker regulations and lower costs to scale up production. The second is an industry relocation effect, in which enterprises facing strict standards, high pollution control costs, and scarce land rationally move capacity to lenient jurisdictions. The study also notes that the spatial network represents statistical correlations of spillover effects rather than physically tracked livestock flows, and that the use of IPCC Tier 1 emission factors may underrepresent mitigation from specific technologies such as anaerobic digesters, limitations the authors flag alongside the measurement error in policy-intensity construction.
The policy implications are pointed. The authors argue that fragmented, purely local spatial squeeze approaches shift environmental costs onto peripheral regions with weaker governance and technology, and they advocate a shift toward regionally coordinated governance centered on efficiency improvement: ecological compensation, joint technology investment, shared industrial parks, and the deliberate export of mature large-scale farming models and waste-utilization technology from the delta and western Guangdong to the north and east. Whether such coordination can convert involuntary capacity reallocation into genuine green upgrading remains an open question, but the Guangdong evidence makes clear that without it, local emission victories may simply be accounting entries in a larger, leakier ledger.
Subject of Research: Spatial leakage of livestock carbon emissions under China's livestock-free zone policies
Article Title: Spatial transfer and policy leakage of livestock carbon emissions driven by local livestock-free zone policies
Article References: Ye, Z., Yang, X., Cao, Y., Liang, Q., Ma, A.-J., & Xu, M. (2026). Spatial transfer and policy leakage of livestock carbon emissions driven by local livestock-free zone policies. Carbon Research, 5(1), Article 51. https://doi.org/10.1007/s44246-026-00283-3
Image Credits: AI Generated
DOI: 10.1007/s44246-026-00283-3
Keywords: carbon emissions, livestock farming, policy leakage, Pollution Haven Hypothesis, Guangdong, spatial Durbin model, difference-in-differences, methane, nitrous oxide, regional governance, spatial spillover, China
News Source: William Thompson. (October 29, 2026). Livestock-Free Zone Policy Cuts Local Emissions but Leaks 3.14 Times More Carbon Next Door. Scienmag.
