Finding pathways to synergistic development of Sustainable Development Goals in China

The study addresses how to effectively monitor and assess sustainable development progress, focusing on the limitations of aggregated measures like the SDG index that mask where improvements are needed. Although the UN SDGs and global indicator framework provide a common standard, sub-national applicability is limited, necessitating localized indicators. Existing national voluntary reviews are largely narrative and lack quantitative guidance for prioritization. The SDG index enables cross-region comparison but suffers from offsetting effects: improvements in some goals can mask deterioration in others, and regional differences complicate interpretation. There is a lack of detailed studies quantifying these offsetting effects and their spatial differences, as well as guidance on leveraging SDG interactions to overcome them. The study focuses on China’s provinces (31, excluding Hong Kong, Macao, Taiwan) after 2015 to: (1) characterize changes in SDGs and offsetting effects; (2) identify spatial similarities and differences in SDG changes; and (3) determine priority SDGs whose promotion would facilitate synergistic progress. Given China’s global importance for achieving SDGs and ongoing national initiatives, clarifying offsetting effects and pathways to mitigate them can inform more effective, timely sustainability policies.

Indicator framework and data: Constructed a provincial-level SDG indicator system for China referencing the UN global indicator framework and literature, applying principles of policy relevance, universal applicability, indicativeness, timeliness, and statistical reliability. The framework includes 88 indicators mapped to 71 targets across 16 SDGs (SDG14 excluded due to non-comparability across inland provinces). Primary data sources were 19 official statistical yearbooks and sectoral datasets (e.g., China Statistical Yearbook, Environmental and Health Statistics Yearbooks). Historical data from 1990 (or earliest availability) to 2018 were compiled to set baselines and targets. Scoring and aggregation: Raw indicator values were normalized to a 0–100 scale using defined baseline and target values (see Supplementary Table S2). Indicator scores were averaged to target scores, which were averaged to SDG scores; the SDG index is the arithmetic mean of SDG scores. Changes from 2015 to 2018 were quantified. Offsetting effects were assessed by counting, at national and provincial levels, the number of SDGs (and targets within SDGs) with increasing or decreasing scores. Statistical tests: Provincial SDG index differences in 2018 were tested via one-way ANOVA, with least significant difference (LSD) post hoc comparisons. Changes from 2015 to 2018 were tested using Mann–Whitney U tests. Analyses were implemented in R v3.6.1 (agricolae and tableone packages). Spatial variation analysis: Hierarchical clustering on provincial SDG score changes (2015–2018) used Euclidean distances and Ward’s method (hclust in R). The optimal number of clusters was determined by the gap statistic (clusGap, cluster package), yielding four provincial groups. Interaction analysis: Multiple Factor Analysis (MFA; FactoMineR::mfa) assessed interactions among three SDG categories per a systematic classification framework: Essential Needs (SDG2, 6, 7, 15), Objectives (SDG1, 3, 4, 5, 8, 10, 16), and Governance (SDG9, 11, 12, 13, 17). Cross-sectional provincial SDG scores for 2015 and 2018 were used to compute RV coefficients (0–1) between categories, with permutation tests for significance. Partial axes plots from MFA informed interaction directionality (acute: synergy; obtuse: trade-off; orthogonal: none). Pearson correlations among SDG scores within and between categories further characterized synergies (r ≥ 0.5), weak synergies (0 < r < 0.5), weak trade-offs (-0.5 < r < 0), and trade-offs (r ≤ -0.5).

- Overall progress: China’s SDG index was 65.4 in 2018, a marginal, non-significant improvement over 2015 (P > 0.05). Beijing ranked highest at 79.3; Gansu lowest at 57.7. Eight provinces were above the national mean; 22 below. Few provinces showed statistically significant score changes 2015–2018. - Best/worst-performing SDGs: SDG13 (Climate Action) performed best nationally; SDG15 (Life on Land) progressed least. Ningxia performed particularly poorly on SDG12 (Responsible Consumption and Production). - Offsetting effects across SDGs: A national decline in an SDG typically corresponded to declines in more than 15 provinces. Conversely, national increases could mask declines in many provinces. From 2015 to 2018, SDG2 (Zero Hunger), SDG10 (Reduced Inequalities), and SDG12 decreased nationally and in more than 15 provinces (changes not statistically significant). SDG15, SDG16, and SDG17 increased nationally, yet each decreased in more than 15 provinces. - Offsetting within SDGs (targets): Despite SDG3 (Health) increasing in all provinces, Target 3.2 (child mortality) and 3.3 (infectious diseases) declined in 24 and 18 provinces, respectively; gains were driven by Targets 3.c (health worker density), 3.8 (basic health services coverage), and 3.6 (traffic fatalities). SDG9 (Industry, Innovation, Infrastructure) increased nationwide, but Targets 9.2 (industrial value added of GDP) and 9.a (investment in pollution control share of GDP) declined in 24 and 17 provinces; improvements in 9.c (mobile network coverage) largely drove SDG9 gains. - Spatial variation and clustering: Eastern provinces had higher 2018 SDG index levels yet more SDGs with declining scores (e.g., Zhejiang ~7 declines, Jiangsu ~8). Northwestern provinces had lower index levels but fewer declines (e.g., Qinghai only SDG15 and SDG17 declined). Some provinces saw index declines due to large drops in specific SDGs despite more goals improving (e.g., Jiangsu SDG16 -17 points; Shandong SDG12 -18.7; Shanghai SDG2 -25.7). Jilin’s declines were offset by >8-point average increases in SDG3, SDG5, and SDG13. Hierarchical clustering grouped provinces into four clusters: Groups I–III commonly had declines in SDG12, SDG15, SDG16, with group-specific additional declines (Group I: SDG2, SDG8; Group II: SDG10; Group III: sometimes SDG6, SDG16). Group IV (Tibet) showed a pronounced decline in SDG7. Increases were common in SDG1, SDG3, SDG4, SDG5, SDG9, SDG11, with magnitude varying by group; Tibet showed large gains in SDG6 and SDG16. - SDG interactions (synergies/trade-offs): In 2015, positive correlations (synergies) existed among the three SDG categories: RV=0.402 (Essential Needs–Objectives), RV=0.307 (Governance–Essential Needs), RV=0.538 (Governance–Objectives), with no strong trade-offs and only some weak trade-offs (notably between SDG7 and Objectives). SDG2 and SDG6 strongly shaped Essential Needs; SDG12 and SDG17 shaped Governance. By 2018, synergies weakened: Governance–Objectives remained significant (RV=0.396, P<0.01), while Governance–Essential Needs (RV=0.109) and Essential Needs–Objectives (RV=0.098) were not significant. Weak trade-offs increased, especially involving SDG7 and SDG15 versus Objectives. SDG7 became a dominant influencer within Essential Needs; SDG12 remained central within Governance. - Priority goals: To mitigate offsetting effects and enhance synergistic progress, prioritizing SDG7 (Affordable and Clean Energy) and SDG12 (Responsible Consumption and Production) is recommended. SDG7 progress was uneven across provinces; SDG12 declined in most provinces and its drop weakened Governance’s coordinating role.

The study quantifies how offsetting effects in aggregated SDG indices can obscure deterioration in specific goals or regions. Even as China’s overall SDG index improved marginally, widespread declines in SDG2 and SDG12 were observed, potentially linked to food safety and unsustainable consumption patterns. Provinces with different geographies and economic levels often face similar development challenges, suggesting that traditional geographic zoning may be suboptimal for policy targeting; cluster-based grouping by SDG change patterns could better support coherent, joint management strategies. Tibet’s distinct challenges, particularly in SDG7, contrast with neighboring Qinghai’s clean energy progress, illustrating the need for tailored, not regionally generalized, interventions. Interactions among SDGs weakened from 2015 to 2018, with more weak trade-offs emerging, especially around SDG7 and SDG15 versus Objectives. Strengthening SDG7 and reversing declines in SDG12 could restore synergies and reduce offsetting effects, fostering more balanced, system-wide progress. These insights underscore the necessity of moving beyond siloed approaches, focusing on both lagging goals within high-scoring regions and declining regions overall, and leveraging cross-goal synergies to guide policy and resource allocation.

An indicator framework for China’s provinces revealed substantial offsetting effects in SDG progress from 2015 to 2018. SDG index changes are strongly influenced by both the number of goals improving or declining and the magnitude of changes in specific SDGs. Provinces across different geographic zones can share similar development dilemmas, suggesting that exploiting similarities (via cluster-informed management) can enhance policy coherence and balanced development. Interaction analyses point to prioritizing SDG12 and SDG7—transforming production and consumption patterns and promoting affordable clean energy—to catalyze synergistic progress across SDGs. Policymakers should focus not only on regions with low or declining SDG index scores, but also on high-scoring regions where several SDGs underperform. Given time-critical global challenges (climate change, pandemics, fragile partnerships), targeted, synergy-seeking strategies are essential to advance towards the 2030 Agenda.

- Indicator framework sensitivity: Results depend on indicator choice; there is no consensus sub-national SDG framework, and many SDGs in this study have fewer than 50% of official indicators due to data constraints. - Data limitations: Reliance on official statistics with varying availability; remote-sensing-derived indicators were not used due to temporal coverage limitations. - Interaction analysis scope: MFA and correlation analyses capture linear associations and are not causal; alternative SDG classifications could change findings. Nonlinear interactions, complex causal feedbacks, and cross-scale effects were beyond the study’s scope. - Generalizability: While the framework is adaptable, contextual differences and evolving data systems may affect transferability; ongoing refinement and integration of additional monitoring tools are needed.