The study addresses the challenge of insufficient green product supply in the market, crucial for achieving green economic growth. Green product innovation (GPI), defined as developing and promoting products balancing environmental protection, resource intensity, and sustainable development, is highlighted as a core element of China's green transformation strategy. The research focuses on regional export-oriented industrial enterprises (REIEs) in China, acknowledging the significant challenges they face in GPI, including high initial costs and stricter environmental regulations. Existing research on GPI drivers focuses on internal organizational influences (resource-based view) and external environmental factors (institutional and stakeholder theories). This study, however, integrates the product space (PS) theory, emphasizing the principle of relatedness and the complexity perspective to understand the interplay of internal and external drivers in driving GPI. The PS theory framework suggests that successful GPI leads to achieving international comparative advantage for newly produced and exported green products. The study aims to investigate three dynamic mechanisms driving GPI: the direct drive of increased green production capacity, the indirect push of provincial-level knowledge capital (green technology and human capital), and the regulatory pull of environmental policy. The paper aims to reveal the composite driving mechanism and inform policy design for supporting green product development and green industries.

The literature review summarizes existing research on the antecedents of green innovation in enterprises, categorizing them into internal organizational influences (based on the natural resource-based view, NRBV) and external environmental factors (institutional and stakeholder theories). Internal drivers discussed include a firm's ability to absorb, understand, and apply green knowledge (absorptive capacity). External drivers primarily focus on environmental regulations (with differing views on their impact) and green finance, highlighting the role of stable funding in supporting green technological innovation. The literature review identifies a gap between theoretical understanding and the practical implementation of green growth policies, particularly concerning the coordination of insufficient green supply and low-level allocation efficiency in the market. The authors contrast their study with previous work by noting the lack of investigation into regional characteristics of GPI and the absence of studies explicitly integrating green complexity and green product-related density within a product space (PS) framework to analyze GPI diversification.

The study utilizes panel data from 2004 to 2018 covering REIEs in 30 provinces in China (excluding Tibet, Hong Kong, Macau, and Taiwan due to data limitations). Green products are identified based on the Comprehensive List of Environmental Goods (CLEG). Data are sourced from the China Customs Database, the China Industrial Enterprises Database, the China Stock Market and Accounting Research (CSMAR) Database, and national and provincial statistical yearbooks. The explained variable, GPI, is proxied by the total amount of green product exports that REIEs newly acquire comparative advantage in each year; a robustness test uses the export volume instead. The explanatory variable, increased green production capacity, is measured using two dimensions from PS theory: green complexity index (GCI) and green product-related density. GCI measures the complex green knowledge stocks accumulated in the production process, while green product-related density measures the association between the products produced by regional enterprises and surrounding products. The driving channels (green technology and human capital) are measured by the number of green patents granted and the level of human capital calculated using the lifetime income method. Moderating variables are green credit (proxied by the ratio of green interest expenditures to total interest expenditures) and environmental regulation (a composite index of industrial wastewater discharge, SO2 emissions, and soot emissions). Control variables include enterprise size, innovation climate, logistics level, FDI, digitization, local government expenditure, and trade openness. The study employs three models: a panel fixed effects regression (direct drive model), an instrumental variable two-stage least squares (IV-2SLS) regression (to isolate the effect of accumulating more complex product-related capabilities), and a three-stage least squares (3SLS) regression (indirect push model to account for the channel effect). A regulatory pull model uses interaction terms between GCI and environmental variables to test moderating effects. A threshold regression model is employed to account for heterogeneity across regions.

The benchmark regression and IV-2SLS results confirm that increased green production capacity directly drives GPI (supporting H1). The 3SLS regression reveals two indirect driving channels: increased green production capacity enhances regional green technology levels (supporting H2a) and human capital levels (supporting H2b), both of which contribute to increased GPI. The study also finds that environmental regulation and green credit have positive moderating effects on the relationship between green production capacity and GPI (supporting H3a and H3b). Threshold regression analysis reveals inverted U-shaped relationships between GPI and green production capacity, FDI, digitization, and trade openness. The proportion of local government expenditure exhibits a diminishing marginal utility effect on GPI. Robustness checks using alternative explained variables, additional control variables, outlier removal, and addressing endogenous reciprocal causation yielded consistent results. The findings are visualized in a figure depicting the moderating roles of environmental regulation and green credit on the relationship between increased green production capacity and GPI.

The findings confirm the direct and indirect effects of increasing green production capacity on GPI. The results support the PS theory's principle of relatedness and the complexity perspective by showing how the accumulation of green knowledge and technology fosters innovation. The positive moderating roles of environmental regulation and green credit highlight the importance of supportive institutional environments in stimulating GPI. The inverted U-shaped relationships suggest potential diminishing returns from excessive investment in certain factors, suggesting optimized levels are necessary for maximizing GPI. The diminishing marginal utility of local government expenditure highlights the need for targeted, differentiated policies considering regional characteristics.

The study makes significant contributions by deconstructing the composite driving mechanism of GPI among REIEs in China, incorporating PS theory, and investigating regional heterogeneity. The findings provide valuable insights for both enterprise managers (emphasizing the development of green capabilities and strategic product choices) and policymakers (highlighting the importance of supportive policies, including environmental regulations and green credit, while recognizing potential diminishing returns from excessive spending). Future research could explore the impact of different types of green credit or environmental regulations and use more granular data to investigate at a firm level.

The study's limitations include the data availability constraints influencing the choice of variables and sample size, particularly the limited access to post-2016 enterprise-level data in the China Customs Database. The use of proxy variables for complex constructs like GPI and green production capacity might also introduce measurement error. Future research could benefit from access to more comprehensive and granular data, potentially using alternative data sources.