Do workers benefit from economic upgrading in the Pearl River Delta, China?

The study investigates whether and how economic upgrading (EU) in China’s Pearl River Delta (PRD) translates into social upgrading (SU) for workers. Motivated by gaps in the global value chain (GVC) literature—such as a focus on single chains rather than geographic regions, limited control for confounders, emphasis on measurable standards over enabling rights, and under-examination of the state’s role—the authors use PRD as a regional case where EU has been pronounced since the 2000s. The research question centers on whether workers have benefited from EU (in wages, social insurance, and voice) and through what mechanisms, particularly the role of state governance in balancing facilitation, regulation, and redistribution. The study posits that in China, unlike some Western contexts where SU often lags EU, the two have proceeded in tandem due to strong public governance, with greater improvements in measurable standards than enabling rights.

The review differentiates EU—improving processes, products, functions, or shifting chains—from SU—enhancing workers’ rights and employment quality. SU comprises measurable standards (e.g., wages, social insurance, working time) and enabling rights (e.g., freedom of association, collective bargaining, voice). Prior GVC studies often find EU does not automatically produce SU, with divergent outcomes by employment status and frequent lagging of SU. Cluster-focused work on tertiarization highlights loss of middle-income manufacturing jobs and polarization in services. Governance is central in mediating EU–SU linkages, spanning public and private actors, functions (facilitation, regulation, redistribution), and scales (international to firm). While Western analyses emphasize trade unions and private governance (CSR), in China, unions are subordinated to the party-state and enterprise management, making public governance especially influential. The review calls for closer examination of the state’s mechanisms in shaping EU–SU outcomes and inclusion of enabling rights and worker-level confounders in analysis.

Data sources: (1) 2012 and 2014 China Labor-force Dynamic Survey (CLDS) for the PRD; (2) semi-structured interviews in 2016 with 7 scholars, 9 employers, 7 officials, and 47 rural migrant workers (RMWs) across Shenzhen (tertiarization) and Foshan (reindustrialization). Sampling for interviews used convenience and snowball approaches in selected urban villages in each city; interviews lasted 1–2 hours. Quantitative design: Dependent variables capture SU: wage (continuous, 2014 deflated to real via CPI), social insurance (medical and pension; dummies), and voice (self-reported 1–5 scale). Independent variable EU equals 0 in 2012 and 1 in 2014, reflecting the documented EU progression (GDP per capita, GDP density, and patents all increased 2012–2014). Control variables: gender, age groups, education, enterprise ownership, industry (secondary, low-end service, high-end service), hukou status (urban, local rural migrant worker—LRMW, outgoing rural migrant worker—ORMW), contract, skill certificate, and trade union membership. Models: OLS for wage and voice; logistic regressions for medical and pension insurance. Sample sizes used in regressions: wage (n=1677), medical insurance (n=1712), pension insurance (n=1712), voice (n=1571). Qualitative design: Interviews triangulate and interpret quantitative findings, probing mechanisms of state regulation, employer responses, and worker experiences of wages/insurance/voice during EU.

Economic upgrading occurred in PRD from 2012–2014: GDP per capita rose from 84,172 to 95,489 yuan (~13%), GDP per km² from 8,710 to 10,023 yuan (~15%), and patents per 10,000 persons from 23.89 to 27.44 (~14%). Regression results (Table 6) controlling for confounders: - EU effect: wage +5,471 yuan (p<0.001); higher odds of medical insurance (OR=1.57, p<0.001) and pension insurance (OR=1.59, p<0.001); modest improvement in voice (coef=0.085, p=0.052). - Education: relative to college+, primary or below had lower wages (−23,738 yuan, p<0.001), lower odds of medical (OR=0.27, p<0.001) and pension (OR=0.32, p=0.001), and lower voice (−0.33, p=0.002). Secondary school also had lower wage and voice. - Age: older groups earned more and had higher pension coverage odds; e.g., 40–49 wage +7,754 yuan (p<0.001), pension OR=2.08 (p<0.001). - Ownership: vs joint venture/foreign, private and individual-owned firms paid less and had lower insurance coverage odds; collective enterprises showed especially low odds for insurance. - Industry: low-end and high-end services associated with higher voice vs secondary industry (coef 0.128 and 0.169, both p<0.05); wages modestly higher in low-end services (p=0.070). - Hukou: LRMWs had lower wages (−4,511 yuan, p=0.013) and lower voice (−0.121, p=0.033) vs ORMWs; urban workers had much higher odds of medical (OR=1.81, p=0.003) and pension (OR=2.24, p<0.001). - Contracts, skill certificates, and union membership increased odds of insurance; skill certificates and union membership also raised wages. Direct comparisons (Table 5) show broad increases 2012–2014 in real wages across groups (e.g., males +6,967 yuan; females +6,652 yuan; high-end services +15,480 yuan), and significant gains in medical and pension coverage for many groups, but smaller and more uneven changes in voice. Qualitative insights: Employers reported increased labor costs due to stricter state enforcement on wages and insurance, with small firms facing bankruptcy or relocation. Workers prioritized wages over insurance; local states applied flexible enforcement to balance economic conditions with labor protection.

Findings indicate that in the PRD, EU has been associated with SU for workers, contrasting with many GVC and cluster studies in developing and Western contexts where SU often lags. Public governance—via the state’s roles in facilitation, regulation, and redistribution—has been central in synchronizing EU and SU. Improvements are stronger in measurable standards (wages, social insurance) than in enabling rights (voice). Mechanistically, frequent minimum wage increases and stricter enforcement boosted low-wage earnings, while social insurance enforcement has been more flexible to avoid harming small enterprises and employment, leading to more moderate gains. Enforcement intensity varies spatially and by firm size, constituting “flexible governance”: stricter where EU performance is stronger and for larger firms. Sectoral dynamics matter: manufacturing upgrading can reduce labor demand and compress worker bargaining power, while service sectors show greater improvements in perceived voice. Education and hukou stratifications persist, shaping differential access to SU benefits.

The study shows that during 2012–2014 in the PRD, EU coincided with SU improvements in workers’ wages, social insurance coverage, and, to a lesser extent, voice. The state’s active public governance balanced facilitation, regulation, and redistribution, enabling EU and SU to proceed in tandem. Measurable standards improved more than enabling rights, reflecting policy priorities and enforcement strategies. Local governments tailored enforcement to economic conditions and firm characteristics, contributing to a positive EU–SU linkage. Future research should further examine workers’ agency and participation in SU as social awareness increases, and explore how governance configurations shape enabling rights over longer horizons and in other regions and sectors.

The analysis is confined to the PRD and the 2012–2014 period, limiting generalizability and longer-term inference. The EU treatment is operationalized as a time indicator (2014 vs. 2012), so causal attribution is constrained by observational design. The voice measure is a subjective self-report (1–5 scale). Interview data were collected in two cities (Shenzhen and Foshan) using convenience and snowball sampling in urban villages, which may introduce selection bias and limit representativeness. Some regression subsamples differ (e.g., n=1571 for voice), and certain subgroup changes are statistically insignificant, indicating heterogeneity in effects.