Picking winners in strategic emerging industries using government subsidies in China: the role of market power

Since 2008, China’s industrial strategy has increasingly emphasized fostering innovative firms in strategic emerging industries to reach the global innovation frontier. While government subsidies are widely used to facilitate innovation and competitiveness, their effectiveness may depend on firms’ market power in upstream (buyer power) and downstream (seller power) markets. Existing literature on whether public R&D complements or crowds out private R&D is inconclusive, particularly in developing contexts, and evidence on labour productivity effects is scarce. This paper investigates how government subsidies affect firm labour productivity in China’s strategic emerging industries and how this relationship is moderated by market power. The study asks whether market power on the seller and buyer sides weakens or strengthens the productivity impact of subsidies. Contributions include: (1) clarifying the role of subsidies under varying market power conditions beyond a pure market-failure rationale; (2) assessing whether subsidies provide sufficient incentives to improve performance when market power is present; and (3) extending the analysis to emerging industries in an emerging economy transitioning toward greater competition.

The review highlights three strands: (1) Innovative firms and labour productivity: Building on the CDM framework, innovation is linked to higher productivity through new technology adoption and human capital accumulation. Competitive industries tend to yield greater productivity advantages for innovators, with concentrated industries showing smaller gains, underscoring market power as a confounding factor. (2) Governmental subsidies and private R&D: Market-failure arguments justify subsidies due to knowledge spillovers and financing frictions. Evidence on additionality versus crowding-out is mixed and context-dependent. In China’s transition toward more open markets, subsidies have supported innovation, sometimes showing inverted U-shaped effects on innovation investment; however, concerns remain regarding adverse selection and crowding-out. (3) Market power and innovation: Buyer power can stifle supplier innovation incentives; the broader implications for subsidy effectiveness under varying market structures remain underexplored. The study extends this literature by examining how buyer and seller power moderate the subsidy–productivity link for Chinese firms in strategic emerging industries.

Empirical model: Starting from a Cobb–Douglas production function Q = A K^β L^γ, total factor productivity A depends on private R&D investment and firm age. After dividing by L and log-transforming, the labour productivity model includes R&D intensity, capital intensity, labour, age, and controls. The panel specification incorporates government subsidies, buyer power (BP), seller power (SP), their interactions with subsidies, industry and year effects, firm nature, and province dummies. A random-effects estimator is used, with endogeneity of R&D/L addressed via two-stage least squares (2SLS), instrumenting R&D/L with its one-year lag (L.R&D/L). Hausman tests confirm endogeneity; weak-instrument tests are performed. Estimation employs Baltagi’s EC2SLS random-effects approach; robustness is checked with a two-stage least-squares first-differenced estimator. Variables: Dependent variable is labour productivity (Q/L) measured as log(1 + operating income per employee). Key regressors: R&D intensity (R&D/L), government subsidy (Subsidy/L), buyer power (BP), and seller power (SP). BP is derived from the share of total purchases from the top five suppliers (S5), normalized to [0,1]; higher BP indicates greater buyer power. SP is derived from the share of total sales to the top five customers (C5), similarly normalized; higher SP indicates greater seller power. Controls include capital intensity (K/L, with K approximated as 10% of total assets), labour (L), firm age, industry dummies (strategic sectors per China’s 2018 classification), year dummies, firm nature, and province dummies. Data: Panel of Chinese publicly listed manufacturing firms in strategic emerging industries, 2006–2019. After exclusions (e.g., missing employment), final sample comprises 10,835 firm-year observations covering 1,392 companies across sectors (new materials, next-generation IT, high-end equipment, biotech, energy conservation and environmental protection, renewable energy, electric cars, digital creative). Descriptive statistics indicate mean R&D/L of 1.280 (SD 0.765) and mean Q/L of 4.340 (SD 0.788).

- R&D intensity and productivity: Across models, R&D/L positively and significantly affects labour productivity (coefficients ~0.153–0.157, p<0.05). Capital intensity (K/L) is strongly positive (~0.673, p<0.05). Firm age and size (L) show positive associations. - Government subsidies: Subsidies positively relate to labour productivity; in the full interaction model, the direct effect of Subsidy is 0.012 (p<0.05). - Market power direct effects: Seller power (SP) negatively relates to labour productivity (e.g., Model 3: −0.060, p<0.05), indicating firms with high seller power exhibit lower labour productivity. Buyer power (BP) shows a small, statistically weaker positive direct effect in the interacted model. - Moderating effects of market power on subsidies: SP*Subsidy is positive (0.025, p<0.05), indicating subsidies’ positive effect on productivity is amplified when seller power is higher. BP*Subsidy is negative (−0.048, p<0.05), indicating subsidies’ positive effect on productivity is weakened when buyer power is higher. - Endogeneity and instrument validity: Hausman tests for endogeneity of R&D/L yield Chi2 = 56.10, 71.59, and 106.79 (all p<0.05). First-stage partial R2 ≈ 0.796; minimum eigenvalue statistics ≈ 15469–15239, rejecting weak instrument concerns. - Model fit: R-squared ≈ 0.505–0.506; Wald Chi2 ≈ 8,224–8,304 (p<0.001). - Robustness: First-differenced 2SLS yields consistent signs and significance: SP*Subsidy ≈ 0.025 (p<0.05), BP*Subsidy ≈ −0.026 (p<0.05); ΔSP is negative and significant; other controls behave similarly. - Heterogeneity: Effects of buyer and seller power and their interactions with subsidies are significant for domestic firms but not for foreign-invested firms, indicating market power moderates subsidy effectiveness primarily among domestic enterprises.

The findings show that while subsidies generally enhance labour productivity, their effectiveness is contingent on firms’ market power positions. High seller power is associated with lower productivity, yet it strengthens the productivity gains derived from subsidies, suggesting that subsidies can partially offset adverse productivity correlates of seller concentration by enabling firms to translate innovation inputs into output more effectively. Conversely, high buyer power undermines the productivity benefits of subsidies, likely reflecting weaker supplier incentives and reduced rent from innovation under powerful buyers. These results address the core question by demonstrating that market structure is a critical moderator of subsidy impacts. For policymakers, this implies that uniform subsidy schemes may yield uneven outcomes across market environments; subsidy design should account for firms’ upstream and downstream power to ensure additionality and avoid unintended efficiency losses.

This study demonstrates that in China’s strategic emerging industries, (1) innovation input (R&D/L) is positively associated with labour productivity; (2) seller market power correlates with lower productivity; and (3) government subsidies improve productivity but are differentially effective depending on market power—enhanced under higher seller power and weakened under higher buyer power. The effects are more pronounced among domestic firms. Policy implications include designing subsidies with attention to firms’ market power positions and complementing subsidies with pro-competitive measures to mitigate adverse market-power effects. Future research should incorporate SMEs and start-ups, explore alternative measures of market power and subsidy design, and utilize expanded post-2020 datasets to assess dynamics under evolving market and policy conditions.

- Sample limited to publicly listed firms in strategic emerging industries; SMEs and start-ups are underrepresented, limiting generalizability. - Observation window ends in 2019 to avoid COVID-19 data distortions; post-2020 dynamics are not captured. - Market power measures (based on top-5 supplier/customer concentration) may not fully reflect broader competitive dynamics or informal relationships. - Despite instrumentation for R&D/L, residual endogeneity or measurement error cannot be entirely ruled out. - Heterogeneity effects for foreign-invested firms are less clear due to smaller subsample size and weaker statistical power.