The increasing global economic integration highlights the significance of free trade in reducing unemployment, expanding markets, and fostering international cooperation. However, free trade also presents challenges due to imbalances in industrial structure and benefit distribution. Free Trade Zones (FTZs), as voluntary agreements among countries or regions, aim to address these issues by promoting trade liberalization and reducing trade barriers. Examples of prominent FTZs include EFTA, NAFTA, and AFTA. China, undergoing a period of economic transformation and facing challenges such as financing constraints and slow development of some enterprises, established an FTZ policy to improve its openness, promote industrial upgrading, and enhance enterprise innovation. This study aims to investigate the impact of China's FTZ policy on firms' innovation performance, focusing on the mechanisms through which this impact occurs and considering potential moderating factors. The study leverages a quasi-natural experiment, utilizing data on FTZ establishment and patent data for listed Chinese companies from 2012 to 2020, to empirically test the effects of FTZs on firms' innovation performance. The study's contributions include developing an analytical framework based on dynamic capability and competitive advantage theories, employing a time-varying DID model with robustness checks, analyzing mediating mechanisms (financing constraints and industrial agglomeration), addressing endogeneity through instrumental variables and propensity score matching, and examining moderating effects and firm heterogeneity.

Existing literature predominantly focuses on the macro-level impact of FTZs, with limited examination of their role in driving firm-level technological innovation. While some research suggests FTZs reduce transaction costs and increase R&D investment, others highlight increased competition pressure on Chinese firms. This study addresses gaps in the literature by focusing on the impact of FTZs on firm innovation performance, employing a rigorous methodological approach, investigating mediating mechanisms (financing constraints and industrial agglomeration), and examining moderating effects and firm heterogeneity. Studies typically focus on the promotion and macro impact of FTZs but overlook the detailed effects on firms' technological innovation. Furthermore, the phased establishment of China's FTZs presents challenges in directly measuring the policy's impact on firm innovation. Finally, existing research does not fully address the mediating roles of industrial agglomeration and financing constraints, nor the moderating effects of economic development level, government subsidies, and firm characteristics.

This study uses a quasi-natural experiment design, leveraging the staggered rollout of China's FTZs. The primary data source includes patent data (total number of patent applications and number of invention patent applications) for listed Chinese companies from the China National Intellectual Property Administration and the CSMAR database. Financial data are sourced from CSMAR and WIND databases, and city/province data from China Statistical Yearbooks. Data are matched based on firms' registered locations. The sample covers 2012-2020, with data cleaning resulting in 10,805 firm-year observations. A time-varying difference-in-differences (DID) model is employed to estimate the impact of FTZs on firms' innovation performance, accounting for company, year, industry, and province fixed effects. Robustness checks include parallel trend tests, placebo tests, high-dimensional fixed effects estimation, and alternative regression models (Tobit, Poisson, and median regression). To address endogeneity, instrumental variable (IV) regression is conducted using industrial SO2 emissions and urban sewage pipe length as instruments, and propensity score matching (PSM) is employed using nearest neighbor, caliper, kernel, and Mahalanobis methods. Finally, moderating effects of economic development level (GDP per capita) and government subsidies, as well as firm heterogeneity (size, ownership, industry, and location), are analyzed.

The baseline regression using a time-varying DID model shows a significant and positive effect of FTZ policy on both the scale (total patent applications) and quality (invention patent applications) of firm innovation. The parallel trend test confirms the validity of the DID model by demonstrating parallel trends between the treatment and control groups before FTZ establishment. The placebo test further strengthens the findings by showing no significant differences between the treatment and control groups in pre-treatment periods. Robustness tests using high-dimensional fixed effects and alternative regression models confirm the baseline results. The mediating effect analysis reveals that the FTZ policy impacts innovation performance by alleviating financing constraints (measured using the SA index) and enhancing industrial agglomeration (measured using the Herfindahl-Hirschman Index (HHI)). Instrumental variable regression, using SO2 emissions and sewage pipe length as instruments, confirms the positive impact of FTZs on innovation even after accounting for endogeneity. Propensity score matching, using various matching methods, yields consistent results, showing significant positive treatment effects. Moderating effects analysis reveals that the impact of FTZs on innovation is stronger in areas with higher economic development (GDP per capita) and government subsidies. Heterogeneity analysis demonstrates that the FTZ policy positively affects innovation performance across various firm characteristics (size, ownership, industry, and location), with stronger effects on large enterprises, non-SOEs, technology-intensive industries, and firms located in the Eastern region.

The findings support the hypothesis that China's FTZ policy has significantly boosted the innovation performance of firms in pilot areas. The mediating roles of financing constraints and industrial agglomeration highlight the importance of these factors in translating policy benefits into enhanced innovation outputs. The moderating effects suggest that the effectiveness of the policy varies across contexts and firm characteristics. The robust results from various econometric techniques, including IV regression and PSM, enhance the credibility of the causal inference. These findings have important implications for policymakers and businesses, suggesting that FTZs can be effective tools for promoting innovation, but their impact is contingent upon contextual factors and firm characteristics.

This study provides strong evidence that China's FTZ policy has substantially enhanced the innovation performance of firms. The mediating roles of financing constraints and industrial agglomeration, along with moderating effects of economic development and government support, highlight the complex interplay of factors involved. Future research could explore the long-term effects of FTZs, delve deeper into the specific mechanisms within financing channels and industry dynamics, and examine the broader societal impacts of this policy, especially in relation to the UN Sustainable Development Goals.

The study focuses on listed companies, potentially limiting the generalizability to privately held firms. The use of patent applications as a proxy for innovation might not fully capture the nuances of innovation activities. While the study addresses endogeneity concerns, unobserved confounders might still exist, although efforts were made to mitigate this through IV and PSM analysis.