Policy Instruments for Inclusive and Sustainable Development: Empirical Insights from China’s Pilot Free Trade Zones

Abstract

Promoting sustainable and balanced economic growth remains a key challenge for developing countries. This study empirically investigates the impact of China’s Pilot Free Trade Zone (PFTZ) on regional economic growth from 2010 to 2023, offering important insights into how targeted policy instruments can contribute to sustainable economic growth. Employing a multiperiod difference-in-differences model and a capital–technology–marketization framework, this study finds that PFTZ implementation has a significant and direct influence on promoting provincial economic growth. The growth effects are primarily driven by improved capital flows and enhanced technological innovation. Notably, these positive effects are more pronounced in central and western Chinese provinces and regions with lagging economic development, indicating that PFTZs can serve as effective tools for reducing regional disparities. These findings provide new empirical evidence regarding the regional heterogeneity of PFTZ policy impacts and offer valuable insights into the design, timing, and spatial targeting of PFTZ initiatives in developing countries seeking to support inclusive and sustainable development across the country.

1. Introduction

In recent years, China’s economic reform has entered a “deep-water zone”—a critical phase requiring structural, institutional, and systemic reforms—where traditional growth models are increasingly inadequate for addressing emerging challenges [1,2]. This calls for new growth engines to sustain economic momentum. As a central government strategy to deepen reform and expand opening-up amid new global conditions, the implementation of China’s Pilot Free Trade Zones (PFTZs) has become vital for regional economic development [3,4]. PFTZs also profoundly impact the country’s broader integration into the global economy as well as sustainable development initiatives [5].

The 2013 launch of Shanghai’s PFTZ ushered in a new phase of China’s opening-up. The zone’s early success provided valuable insights for subsequent PFTZs. Over the past decade, China has established 22 PFTZs through seven rounds of expansion, forming a coordinated national framework characterized by east–west, north–south, and land–sea integration. These PFTZs have refined their developmental positioning and implemented new institutional reforms across trade, investment, data governance, and finance to maximize policy effectiveness. By early 2025, the 22 PFTZs had made significant progress in multiple domains, attracting USD 28.25 billion in actual foreign direct investment (FDI), accounting for 24.3% of the national total [6].

Amid growing international uncertainty and instability in economic globalization, harnessing the economic effects of PFTZs is particularly crucial [7,8,9]. To further explore the sustainability of the economic effects of PFTZs, this study uses per capita gross domestic product (GDP) as the dependent variable; green total factor productivity (GTFP) as the replacement variable for the dependent variable; and indicators like urbanization level, educational level, and infrastructure level as control variables. Overall, this study includes two dependent variables (tested separately), one core explanatory variable, five control variables, and three mediating variables. Most of the data are sourced from the China Statistical Yearbook and other official websites. Based on existing literature, this study proposes three key research questions for further exploration. (1) Is the PFTZ policy a primary driver of provincial economic growth? In many cases, PFTZ implementation coincides with other regional policies such as the Belt and Road Initiative (BRI). Therefore, it is necessary to isolate and evaluate the PFTZs’ independent contribution in regions subject to overlapping policy interventions. (2) Through what mechanisms do PFTZs influence provincial economic growth? While existing studies have primarily focused on the direct relationship between PFTZs and economic performance [10,11], the underlying transmission mechanisms remain underexplored. A deeper understanding of these channels is essential for identifying potential new growth drivers and improving policy design. (3) Do the economic effects of PFTZs vary by region and economic scale? PFTZs are intended to promote balanced and sustainable regional development, but their differential impacts across provinces—particularly between economically advanced eastern regions and less-developed central and western regions—raise concerns regarding potential inequality. Understanding this heterogeneity is crucial for evaluating the inclusive effectiveness of PFTZ policy implementation promoting sustainable economic growth among different regions.

The remainder of this study is structured as follows. Section 2 outlines the theoretical framework and develops our research hypotheses. Section 3 presents the empirical design, including variable definitions, sample selection, and data processing methods. Section 4 details the results of the baseline regressions, robustness tests, heterogeneity analysis, and mediating effect tests. Section 5 concludes with a summary of findings, policy implications, and discussion of the research.

2. Theoretical Framework and Research Hypotheses

PFTZs function as policy laboratories for institutional innovation. Drawing on economic theories such as endogenous growth and institutional economics, this section develops an analytical framework linking PFTZs to capital flows, technological innovation, and market-oriented reforms, from which empirically testable hypotheses are derived.

2.1. Pilot Free Trade Zones and Provincial Economic Growth

As a location-targeted policy tool, PFTZs are intended to stimulate economic growth within associated jurisdictions and in adjacent regions to address regional development imbalances. Previous research has indicated that PFTZs are among the primary drivers of regional economic expansion [12]. From the perspective of factor mobility, PFTZs enhance capital and labor allocation efficiency by optimizing the business environment and facilitating trade and investment flow [13]. From the structural transformation perspective, PFTZs support industrial upgrading and trade diversification, contributing to economic growth [14,15,16]. Overall, PFTZ policies are a comprehensive institutional mechanism that attracts high-quality production factors and aggregates high-end industries, while also improving financial services and optimizing the regulatory environment. These factors collectively enhance investment and trade activity, stimulate technological innovation, and contribute to sustainable regional economic growth [17,18]. Some studies have employed policy evaluation methods such as the difference-in-difference (DID) approach and the synthetic control method to further examine the economic effects of early PFTZs established in different regions. These studies have indicated that although the degree of economic impact varies across zones, PFTZs have indeed had a positive impact on promoting local economic development [19,20]. Furthermore, PFTZs exert positive spillover effects on surrounding regions [21].

Accordingly, this study proposes the following hypothesis.

Hypothesis 1.

Pilot Free Trade Zone implementation significantly promotes provincial economic growth. (That is, the data must be significant at the 10%, 5%, and 1% levels, and the coefficient must be positive.)

2.2. Mechanisms

The endogenous growth theory, as the key guiding theory of this study, posits that sustained economic growth primarily stems from investments in knowledge, human capital, and technological innovation [22]. As practical zones for institutional innovation, PFTZs provide a unique experimental environment for studying how relevant policies, institutions, and measures can activate these endogenous growth drivers [23]. Specifically, PFTZs function as laboratories aimed at overcoming institutional barriers, reducing transaction costs, and fostering an environment conducive to technological innovation, knowledge spillovers, capital accumulation, and market efficiency—all core elements of endogenous growth theory [24]. Therefore, this study establishes a connection between the unique institutional innovations of PFTZs and the elements of endogenous growth theory. Drawing on the theoretical frameworks of Zeng, Barro as well as the core tenets of the neoclassical growth model (i.e., that capital flows are the primary driver of economic growth) [25] and other scholars [26,27,28,29], this study constructs a capital–technology–marketization mechanism model to analyze how PFTZs influence provincial economic growth. The study also refines this framework by identifying appropriate proxy variables and theoretical constructs from the existing literature.

Within this analytical framework, capital, technology, and marketization form a three-dimensional interlocking enhancement system. Specifically, market-oriented reforms reduce institutional costs through measures such as negative list reduction and cross-border data flow pilots [30,31]. This not only attracts cross-border capital to cluster in core technology sectors but also forces institutional innovation driven by capital flow demands. Capital flows inject momentum into technology R&D through foreign funding while benefiting from efficiency gains in cross-border payments enabled by technological innovation [32]. Technological innovation relies on testing scenarios provided by market reforms while reciprocally upgrading market supervision models through its latest achievements [28,33]. This multidirectional interaction creates a deep coupling of “institutional supply-capital empowerment-technological innovation-market upgrading,” ultimately driving systemic leaps in PFTZ economic growth.

2.2.1. Mechanism I: Capital Flow

Capital flows is an important driver of growth in economic theory. From an endogenous growth perspective, efficient capital allocation, particularly towards productive investments and knowledge-based activities, is crucial [34]. PFTZs function as experimental zones for China’s broader strategy of economic openness, facilitating inbound and outbound capital flows that are crucial for advancing regional economic development [35]. Crucially, this facilitation is achieved through targeted institutional innovations, such as aligning with high-standard investment and trade rules, relaxing capital controls, and implementing a “negative list” approach to FDI [36,37]. These institutional changes reduce barriers and uncertainties, thereby enhancing the efficiency of capital flow and its allocation towards potentially more productive uses, consistent with endogenous growth dynamics [38]. Furthermore, capital flows directly benefit regional economies. Inbound capital can spur industry development, improve competitiveness, and increase productivity (especially through support for knowledge-intensive SMEs), while outbound capital helps domestic firms enhance risk resilience and expand their international market share [39]. It is the PFTZ’s institutional innovation that enables this enhanced capital mobility and its potential for growth-enhancing effects.

Accordingly, this study proposes the following hypothesis.

Hypothesis 2.

PFTZs promote provincial economic growth by facilitating capital flow.

2.2.2. Mechanism II: Technological Innovation

The endogenous growth theory places technological progress and knowledge creation at the very heart of sustained growth, emphasizing spillovers, R&D investment, and human capital accumulation [40]. PFTZs serve as catalysts for technological progress by leveraging institutional innovation to establish favorable innovation ecosystems that encourage domestic and foreign firms to invest in R&D and adopt advanced technologies. On the one hand, PFTZs’ institutional innovations (e.g., preferential policies, improved governance, enhanced intellectual property protection) attract foreign enterprises [41]. Their presence generates knowledge spillovers and competitive pressure, encouraging domestic firms to innovate and upgrade. Foreign firms also contribute by transferring knowledge, skills, and management practices to the local ecosystem [42,43]. On the other hand, technological innovation fostered within this institutionally supportive environment directly enhances productivity, product quality, and firm competitiveness, which are key drivers of regional economic development [44].

Accordingly, this study proposes the following hypothesis.

Hypothesis 3.

PFTZs promote provincial economic growth by enhancing technological innovation capacity.

2.2.3. Mechanism III: Market-Oriented Reform

PFTZs are not only zones of economic liberalization but also laboratories for institutional reform aimed at enhancing market efficiency. By streamlining regulations, fostering governance modernization, and transforming the functions of government agencies, PFTZs drive market-oriented reform [45]. Specific initiatives such as simplifying administrative procedures, reducing regulatory burdens, and liberalizing trade and investment regimes are intended to enhance market efficiency, lower transaction costs, and improve the business environment [46]. These institutional changes are “deliberate interventions” aimed at creating a more efficient market within PFTZs, thereby alleviating some of the inefficiencies that hinder endogenous growth. Empirical studies have consistently demonstrated the positive association between such market-oriented institutional reforms and economic performance [47,48,49,50]. For example, Iradian analyzed panel data from 123 countries, concluding that sustained economic growth in transitional economies is closely correlated with successful market-oriented reforms [51].

Accordingly, this study proposes the following hypothesis.

Hypothesis 4.

PFTZs promote provincial economic growth by advancing market-oriented institutional reforms.

3. Research Design

The empirical analysis employs a multiperiod DID estimation strategy, leveraging the phased implementation of PFTZs across provinces in China to explore the impact of PFTZ policy on regional economic sustainability. This section provides a detailed account of model specification, variable construction, and data source analysis to ensure the validity of the research data.

3.1. Model Specification

PFTZ implementation in China serves as a quasi-natural experiment due to the staggered timing of policy implementation across regions. This study employs a multiperiod DID estimation strategy to rigorously evaluate the impact of PFTZs on provincial economic growth. Compared with alternative identification methods, the multiperiod DID approach is more robust in addressing treatment heterogeneity and policy implementation timing, while mitigating endogeneity concerns that arise from time-invariant unobservables [52]. This study defines provinces, municipalities, and autonomous regions that have established PFTZs as the treatment group and non-PFTZs as the control group. This study excludes observations from Tibet, Hong Kong SAR, Macao SAR, and Taiwan Province due to a lack of consistent and reliable data. The final panel dataset includes annual observations from 30 provincial-level administrative regions spanning 2010 to 2023. Referring to the research of Minzhe Du, Chukun Huang, and Liping Liao [53], the baseline estimation model is specified as follows:

$${LnAGDP}_{i,t} = {\alpha }_0 + {\alpha }_1{did}_{i,t} + \beta X_{i,t} + {\gamma }_i + {\delta }_t + {\epsilon }_{i,t}$$

(1)

where i denotes provinces, municipalities, and autonomous regions in China, and t represents the time dimension (2010–2023). LnAGDP is the natural logarithm of per capita GDP as a proxy for regional economic growth. ${did}_{i,t}\left({treat}_i{PFTZ}_{i,t}\right)$ represents the core explanatory variables of this study, of which the variable treat functions as a grouping dummy variable where the treatment group is assigned the value of 1 and the control group is valued at 0. PFTZ is the policy dummy variable, taking the value of 1 for all years from PFTZ establishment onward and 0 otherwise.

The model further controls for a vector of covariates ($X_{i,t}$) encompassing five key regional characteristics of urbanization level (UL), educational level (EL), infrastructure (IL), foreign trade dependence (FTD), and fiscal expenditure (FE). UL captures the degree of urbanization, EL controls for human capital differences; IL measured per capita urban road area to account for variation in infrastructure development; FTD reflects regional economies’ openness to international trade; and FE, defined as the ratio of government budgetary expenditure to GDP, proxies for the degree of government intervention in the economy. To address unobserved heterogeneity, the model includes region (${\mathsf{\gamma }}_{\mathrm{i}}$) and year (${\mathsf{\delta }}_{\mathrm{t}}$) FEs, establishing a two-way FE specification. This approach can mitigate potential omitted variable bias and improve the robustness of the estimated policy effects.

3.2. Variable Selections and Data Sources

3.2.1. Variable Selections

Dependent Variable: Per capita GDP (LnAGDP).

In the literature examining the economic effects of PFTZs, commonly used indicators of economic growth include total factor productivity (TFP), GDP, and per capita GDP [54], among which, GDP and per capita GDP provide a direct and reliable reflection of a country’s or region’s economic development over time [55]. However, GDP is relatively more susceptible to time-specific shocks, which may affect its reliability as a measure of sustained economic growth [5]. In addition, inconsistencies in statistical standards and a lack of complete and comparable time series data across certain regions make it difficult to obtain reliable TFP estimates [56]. Considering these limitations, the study adopts per capita GDP (measured in yuan) as the dependent variable to assess PFTZs’ economic impact on regional development. This study log-transforms the variable to reduce heteroskedasticity and account for scale effects.

Core Explanatory Variable: ${did}_{i,t}$ (dummy variable).

Referencing Wang et al. [57], the study constructs a treat $\times$ PFTZ interaction term, in which treat is a grouping dummy variable (treatment and control groups), where regions that established PFTZs are in the treatment group (assigned the value of 1) and the control group is assigned a value of 0. PFTZ is the dummy variable of PFTZ implementation, i.e., if a region has an officially listed PFTZ in the year, it is assigned a value of 1 in that year and subsequent years, and 0 in the rest of the years. For example, if the Shanghai PFTZwas officially established in 2013, it is assigned a value of 1 for the period from 2013 to 2023.

Control Variables

The model includes the following control variables to account for potential confounders.

(1) Urbanization Level (UL): Urbanization level has a significant influence on regional economic development. The study measures this variable as the proportion of the urban population at year-end (in %). A higher urbanization rate is expected to have a positive effect on local economic growth [58].

(2) Educational Level (EL): The study quantifies this measure as the ratio of students enrolled in general and vocational undergraduate and junior college programs to the total population (in %). Higher educational levels contribute to talent accumulation, narrow urban–rural disparities, and promote economic development [59,60].

(3) Infrastructure Level (IL): Infrastructure is foundational to the smooth operation of socioeconomic activities and strongly correlated with regional economic performance [61,62,63]. This study uses per capita urban road area (in square meters) as a proxy, with data sourced from the China Statistical Yearbook.

(4) Foreign Trade Dependence (FTD): Since China’s reform and opening-up, the nation’s rapid economic growth has largely benefited from an export-led development model. However, overreliance on external markets may adversely affect domestic enterprises’ long-term viability [64]. To quantify trade openness, this study uses the ratio of total import–export trade volume (in CNY 100 million) to regional GDP (in CNY 100 million).

(5) Fiscal Expenditure (FE): Government intervention through fiscal spending can influence the allocation of market resources; however, the extent of such intervention, reflected in fiscal expenditure, and its effectiveness vary considerably across regions, largely due to differences in local governance capacity and policy priorities. Notably, higher fiscal expenditure can indicate weaker budgetary constraints and greater government intervention in the economy [37]. This study measures fiscal expenditure as the ratio of local general public budget expenditure to regional GDP (in CNY 100 million, expressed in percentage) to capture the degree of government intervention.

Mediating Variables

This study incorporates the following mediating variables to explore the potential mechanisms through which PFTZs affect regional economic growth.

(1) Foreign direct investment (LnFDI)–Capital: To quantify the effect of capital flow, this study references Marin and Schnitzer [65], employing the actual use of FDI as a proxy for capital mobility. The FDI data are sourced from the China Statistical Yearbook and the annual statistical bulletins of each region (unit: CNY million), which are log-transformed to address potential skewness and scale effects.

(2) Patent grants (apply)–Technology: There are a variety of indicators for measuring technological innovation, such as R&D investment, patent grants, and high-tech exports. However, due to incomplete data for some indicators and the fact that R&D investment more accurately reflects the resource consumption of the innovation process, this study also refers to the approach from some studies [44,66] and uses the patent grants, which represents output, as a proxy variable. This variable is measured by the number of authorized invention patents (unit: items), scaled by a factor of 10,000 to reduce heteroskedasticity.

(3) Marketization index (market)–Marketization: Market development is widely recognized as a crucial factor in assessing the effectiveness of economic reforms [67,68]. Following the prevailing literature [69,70,71], this study employs the regional marketization index as a proxy variable. The data are obtained from the China Sub-provincial Marketization Index Database, integrating index scores and their conceptual dimensions.

In summary, this study includes one dependent variable, one core explanatory variable, five control variables, and three mediating variables. The specific names, calculation methods, and data sources of these variables are presented in

Table 1. Variable definitions and data sources.

Form	Name	Sign	Calculation Method	Data Sources
Dependent Variable	Per Capita GDP	LnAGDP	Log-transformed per capita GDP	China Statistical Yearbook
Core Explanatory Variable	Dummy Variable	did	Dummy variables (0, 1)	Official website of the Pilot Free Trade Zones in each region
Control Variables	Urbanization Level	UL	Number of urban population at the end of the year in the region/total population of the region	China Statistical Yearbook
	Educational Level	EL	Number of students enrolled in general and vocational undergraduate programs in the region/total population of the region	China Statistical Yearbook
	Infrastructure Level	IL	Urban road space per capita	China Statistical Yearbook
	Foreign Trade Dependence	FTD	Total import and export trade/regional GDP	China Statistical Yearbook
	Fiscal Expenditure	FE	Local general public budget expenditure/regional GDP	China Statistical Yearbook
Mediating Variables	Patent Grants	apply	Number of authorized patents for inventions	China Statistical Yearbook
	Marketization Index	market	Marketization index	USTB
	Foreign Direct Investment	LnFDI	Log-transformed actual utilized foreign direct investment	China Statistical Yearbook, annual local statistical bulletin

Note: The table was compiled by the author (same below).

. Descriptive statistics for the main variables are presented in

Table 2. Descriptive statistics.

Form	Variable	Mean	Standard Deviation	Min	Max
Dependent Variable	LnAGDP	10.897	0.485	9.482	12.207
Core Explanatory Variable	did	0.324	0.469	0	1
Control Variables	UL	60.000	12.327	33.81	89.6
	EL	0.021	0.006	0.008	0.044
	IL	16.568	5.179	4.04	28
	FTD	0.267	0.273	0.007	1.493
	FE	0.256	0.111	0.107	0.758
Mediating Variables	apply	7.275	11.67	0.026	87.221
	market	8.187	2.003	3.36	13.36
	LnFDI	13.352	1.507	9.618	17.949

. The results confirm that the variables contain no outliers and fall within a normal range.

3.2.2. Correlation Analysis

The study performs correlation analysis on the processed sample data, as shown in

Table 3. Results of correlation analysis.

	LnAGDP	did	UL	EL	IL	FTD	FE
LnAGDP	1.000
did	0.555 ***	1.000
UL	0.868 ***	0.448 ***	1.000
EL	0.564 ***	0.432 ***	0.593 ***	1.000
IL	−0.055	−0.028	−0.074	−0.053	1.000
FTD	0.476 ***	0.220 ***	0.643 ***	0.149 ***	−0.024	1.000
FE	−0.385 ***	−0.314 ***	−0.314 ***	−0.336 ***	−0.001	−0.437 ***	1.000

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, to better understand the relationships between the sample data. The correlation analysis reveals a significant correlation between the sample data, and most control variables (e.g., UL, EL, and FTD) have a strong relationship with the regional economic growth (LnAGDP), and most are positively significant at the 1% confidence level.

3.2.3. Multicollinearity Analysis

To further ensure the scientific validity and robustness of our empirical model, the study conducts a multicollinearity analysis using the variance inflation factor (VIF). A VIF value exceeding 10 is commonly considered indicative of problematic multicollinearity, with higher values signifying greater collinearity among variables [72]. As shown in

Table 4. Multiple covariance analysis.

Variable	VIF	1/VIF
UL	3.34	0.299269
FTD	2.48	0.402616
EL	2.11	0.474482
FE	1.48	0.674851
did	1.37	0.728884
IL	1.01	0.993294
Mean VIF	1.97

, all VIF values are well below the critical value of 10, with a mean VIF of 1.97. These results indicate that multicollinearity in the current dataset is not a concern, and the model specifications are appropriate for subsequent empirical analysis.

4. Empirical Analysis

This section systematically assesses the impact of PFTZs on regional economic sustainability through empirical analysis, including benchmark regression analysis, parallel trend test, and mediating effect test, etc. The research results collectively corroborate the growth mechanism and lay the foundation for further policy implications in the following sections.

4.1. Benchmark Regression

This section estimates the model using the specified variables, presenting the results in

Table 5. Benchmark regression results.

	(1)	(2)	(3)	(4)	(5)
	Original Form	Eastern Region	Central-Western Region	Lower Level of Economic Development	Higher Level of Economic Development
	LnAGDP	LnAGDP	LnAGDP	LnAGDP	LnAGDP
did	0.075 ***	0.002	0.159 ***	0.041 *	0.099
	(2.738)	(0.113)	(3.322)	(1.665)	(1.459)
UL	0.036 ***	0.007 *	0.038 ***	0.024 ***	0.032 ***
	(5.840)	(1.861)	(3.958)	(3.249)	(4.203)
EL	−7.459	−33.563 ***	7.156	15.998 **	−19.087
	(−1.225)	(−6.830)	(1.162)	(2.500)	(−1.449)
IL	0.000	0.028 ***	0.000	0.000	0.020
	(0.070)	(5.462)	(1.579)	(1.567)	(1.266)
FTD	−0.338 **	−0.198 ***	−0.100	−0.144	−0.282 **
	(−2.314)	(−2.691)	(−0.548)	(−0.599)	(−2.047)
FE	−0.672 **	−0.295	−0.073	−1.287 ***	0.811
	(−1.971)	(−0.612)	(−0.176)	(−4.524)	(0.821)
_cons	8.790 ***	11.563 ***	8.297 ***	8.998 ***	9.018 ***
	(33.607)	(50.769)	(20.013)	(26.700)	(14.078)
N	420	140	238	210	210
R2	0.922	0.985	0.879	0.954	0.847
province effect	Yes	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

. Column (1) details the baseline estimation of PFTZs’ economic effects. After controlling for a series of covariates and region and year FEs, the coefficient of did is 0.075 and statistically significant at the 1% level. This indicates that PFTZ implementation significantly promotes regional economic growth by approximately 7.5%, empirically supporting Hypothesis 1.

Regarding control variables, the coefficient UL (0.036) is significantly positive, indicating that urbanization enhances regional economic performance. This may be attributable to urbanization’s role in facilitating the freer movement of production factors between urban and rural areas, unlocking local consumption potential and stimulating economic growth. In addition, the coefficients for FTD and FE are also significant at the 5% level, indicating a strong correlation between these factors and per capita GDP.

Columns (2) and (3) of Table 5 present the results of the heterogeneity analysis by geographic location. Compared with the eastern region, PFTZs in central and western provinces exhibit a larger and more significant economic effect, with a coefficient of 0.159. This finding indicates that PFTZ policies are significantly effective in driving economic growth in less-developed inland regions, contributing to more balanced regional development.

Columns (4) and (5) of Table 5 explore economic development heterogeneity. This study classifies regions into two groups according to the average per capita GDP, categorizing the top 50% as economically developed and the bottom 50% as less developed. The results reveal that both groups yield positive coefficients, but the effect is only statistically significant at the 10% level for the less-developed group, while the effect for the other group is insignificant. This indicates that the economic benefits of PFTZs are more pronounced in regions with weaker initial economic conditions, which is likely due to their greater potential for catch-up growth and stronger marginal impact of PFTZ implementation in these areas.

4.2. Parallel Trend Test

Referencing Wang et al. and Feng et al. [7,73], this study employs an event study method to examine whether economic development trends between the treatment and control groups were consistent prior to PFTZ implementation. To do so, this study subdivides the treatment dummy variable into three years before and five years after PFTZ implementation, constructing interaction terms between time dummies and group identifiers to capture the dynamic effects of PFTZs. These interactive terms are introduced into the model for the parallel trend test as follows:

$$\begin{array}{l}{LnAGDP}_{i,t} = {\alpha }_0 + {\alpha }_1{treat}_i \times {{before}^{-3}}_{i,t} + {\alpha }_2{treat}_i \times {{before}^{-2}}_{i,t} +\\ {\alpha }_3{treat}_i \times {{before}^{-1}}_{i,t} + {\theta }_1{treat}_i \times {{PFTZ}^0}_{i,t} + {\theta }_2{treat}_i \times P{{FTZ}^{+1}}_{i,t} +\\ {\theta }_3{treat}_i \times {{PFTZ}^{+2}}_{i,t} + {\theta }_4{treat}_i \times {{PFTZ}^{+3}}_{i,t} + {\theta }_5{treat}_i \times P{{FTZ}^{+4}}_{i,t} +\\ {\theta }_6{treat}_i \times {{PFTZ}^{+5}}_{i,t} + \beta X_{i,t} + {\gamma }_i + {\delta }_t + {\epsilon }_{i,t}\end{array}$$

(2)

where the definitions of ${ LnAGDP}_{i,t},{treat}_i$, and $X_{i,t}$ remain unchanged; ${{before}^{-n}}_{i,t}$ represents each region’s development status before PFTZ implementation (n takes the value of 1–3); and ${{FTZ}^{+m}}_{i,t}$ represents each region’s development status after PFTZ implementation (m takes the value of 0–5).

As illustrated in Figure 1, the parallel trend test includes data from three years before and five years after the policy implementation, with t = 0 representing the year the PFTZ policy was implemented. The estimated coefficients of the treatment effect prior to policy implementation are all close to zero and statistically insignificant, satisfying the key assumption of the parallel trend test. After the policy is enacted, most coefficients become statistically significant and exhibit considerable fluctuations, forming an inverted V-shape. This pattern indicates that although some successful experiences from PFTZ operations have been summarized and promoted nationwide, regional heterogeneity limits the applicability of these experiences. As a result, the PFTZ policy’s effectiveness may vary and is not always significant.

4.3. Placebo Test

To rule out the influence of other random factors, the study conducts a placebo test by randomly reassigning the treatment group (keeping the number of observations unchanged) and the timing of PFTZ implementation. New dummy variables are generated and introduced into the baseline regression, and the process is repeated 500 times to obtain the final placebo test results [43]. As shown in Figure 2, the benchmark regression coefficient from the original model (solid line on the right) does not intersect with any of the coefficients from the 500 simulated regressions. This outcome satisfies the requirements of the placebo test, further demonstrating a strong correlation between regional economic development and PFTZ implementation, and validating the robustness of the baseline model.

4.4. Other Robustness Analyses

4.4.1. Changing the Control Group

Considering the potential time lag in PFTZ policy effects, the study adjusts the group classification for the Xinjiang PFTZ, which was established in 2023. As shown in

Table 6. Exclusion of other factors.

	Change Control Group			Exclusion of Other Policy Factors
	(1)	(2)	(3)	(4)
	LnAGDP	LnAGDP	LnAGDP	LnAGDP
did	0.075 ***	0.076 ***	0.075 ***	0.075 ***
	(2.739)	(2.675)	(2.738)	(2.738)
DID	/	/	/	−0.022
	/	/	/	(−0.274)
UL	0.036 ***	0.036 ***	0.036 ***	0.036 ***
	(5.846)	(5.775)	(5.840)	(5.840)
EL	−7.241	−7.697	−7.459	−7.459
	(−1.188)	(−1.227)	(−1.225)	(−1.225)
IL	0.000	0.000	0.000	0.000
	(0.075)	(0.074)	(0.070)	(0.070)
FTD	−0.339 **	−0.346 **	−0.338 **	−0.338 **
	(−2.323)	(−2.297)	(−2.314)	(−2.314)
FE	−0.659 *	−0.664 *	−0.672 **	−0.672 **
	(−1.936)	(−1.871)	(−1.971)	(−1.971)
_cons	8.782 ***	8.770 ***	8.790 ***	8.790 ***
	(33.538)	(32.336)	(33.607)	(33.607)
N	420	406	420	420
R2	0.922	0.921	0.922	0.922
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, column (3) presents the original benchmark regression including the Xinjiang PFTZ in the experimental group, column (2) shows the regression results excluding Xinjiang from the sample, and column (1) includes Xinjiang in the control group. Across all three scenarios, the core explanatory variable remains positively significant at the 1% confidence level. This confirms that the baseline conclusions are robust and PFTZ implementation has a significantly positive impact on regional economic development.

4.4.2. Controlling for Other Policy Effects

Since some provinces may be simultaneously influenced by multiple overlapping policies, this study conducts a robustness test by excluding potential interference from other policies. Using the BRI as an example and referencing Feng et al. [74], this study constructs a BRI dummy variable: provinces containing a key BRI node city are coded as 1 from 2013 onward and 0 otherwise. This BRI-related DID variable is then incorporated into the baseline model (Equation (1)) to re-estimate the policy effects [75] as follows:

$${LnAGDP}_{i,t} = {\theta }_0 + {\theta }_1{did}_{i,t} + {\theta }_2{DID}_{i,t} + \beta X_{i,t} + {\gamma }_i + {\delta }_t + {\epsilon }_{i,t}$$

(3)

where ${DID}_{i,t}$ represents the BRI policy, and all other variables remain consistent with the definitions in Equation (1). The regression results are presented in Table 6. The coefficient of the core explanatory variable for the PFTZ policy is 0.075 and statistically significant at the 1% confidence level after incorporating the new BRI dummy variable. The BRI dummy variable itself is not significant, indicating that PFTZ implementation is a primary driving factor for regional economic development, and the observed policy effect is not confounded by the BRI’s influence.

4.4.3. Replace Dependent Variable

To investigate the sustainable impact of PFTZ policy on economic growth while ensuring the robustness of the regression results, this study draws on the views of other scholars [76] and uses green total factor productivity as a proxy variable for sustainable economic growth. Green total factor productivity (GTFP) is one of the key indicators for measuring economic development. It is based on total factor productivity and incorporates green energy factors into the statistical scope, such as industrial wastewater discharge, exhaust emissions, and energy consumption, thereby establishing a comprehensive evaluation system. In a word, this study follows Liu’s approach and uses the SBM-GML model to calculate the data [77], with the specific indicator data shown in

Table 7. GTFP definition.

Category	Name	Specific Data
Factor Inputs	Labor Input	Provincial Employment Numbers
	Capital Input	Actual Capital Stock
	Energy Consumption	Total Energy Consumption
Desirable Outputs	Economic Output	GDP
Undesirable Outputs	Industrial Wastewater Discharge	Industrial Wastewater Discharge Volume
	Industrial Gas Emissions	Industrial SO2 Emissions
	Industrial Solid Waste Discharge	Industrial Solid Waste Discharge Volume

below:

Replacing the LnAGDP in Formula (1) with GTFP data, the benchmark regression results are shown in

Table 8. Benchmark regression results-GTFP.

	(1)	(2)	(3)
	Original Form	Eastern Region	Central-Western Region
	GTFP	GTFP	GTFP
did	0.025 **	0.004	0.025 *
	(2.044)	(0.183)	(1.803)
UL	−0.001	0.005	0.001
	(−0.388)	(1.139)	(0.463)
EL	4.439	17.367	−1.718
	(1.598)	(1.632)	(−0.773)
IL	0.000	−0.006	0.000 *
	(0.319)	(−1.158)	(1.983)
FTD	−0.191 ***	−0.367 ***	0.046
	(−2.877)	(−4.859)	(0.607)
FE	−0.130	−1.258 *	0.105
	(−0.836)	(−2.248)	(0.774)
_cons	1.034 ***	0.542	0.952 ***
	(8.668)	(1.676)	(9.769)
N	420	140	238
R2	0.331	0.442	0.337
province effect	Yes	Yes	Yes
time effect	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

. It can be seen that, after controlling for a series of variables, the did is positively significant at the 5% level. This suggests that the establishment of PFTZs can also enhance green total factor productivity. Combined with other empirical results (when LnAGDP is the dependent variable, the model is also positively significant), this further indicates that PFTZ policy can indeed promote sustainable economic development. Additionally, based on this, this study also conducted heterogeneity tests for this variable across eastern, central, and western regions. The results show that the impact of PFTZ policy on central and western regions is more positive and significant. This may be because eastern regions are mostly coastal, economically developed, and open, with a greater emphasis on environmental protection and governance, while central and western regions are mostly inland, making the impact of PFTZ policy more pronounced, helping them expand their openness and better address current environmental challenges [76].

4.5. Mediating Effect Test

As previously described, this study employs a capital–technology–marketization framework and applies a mediating effect test to explore whether these factors serve as mechanisms through which the PFTZ exerts economic impact. Referencing Jiang [78], the study constructs the following models, where the core explanatory variables and control variables remain consistent with those used in previous equations. Specifically, ${LnFDI}_{i,t}$, ${apply}_{i,t}$, and ${market}_{i,t}$, respectively, denote capital flow, technological innovation capacity, and market-oriented reforms.

$${LnFDI}_{i,t}={\phi }_0+{\phi }_1{treat}_i{PFTZ}_{i,t}+\beta X_{i,t}+{\gamma }_i+{\delta }_t+{\epsilon }_{i,t}$$

(4)

$${apply}_{i,t}={\phi }_0+{\phi }_1{treat}_i{PFTZ}_{i,t}+\beta X_{i,t}+{\gamma }_i+{\delta }_t+{\epsilon }_{i,t}$$

(5)

$${market}_{i,t}={\phi }_0+{\phi }_1{treat}_i{PFTZ}_{i,t}+\beta X_{i,t}+{\gamma }_i+{\delta }_t+{\epsilon }_{i,t}$$

(6)

According to the mediating effect model proposed by Jiang [78], once it is established that the three elements of capital, technology, and marketization are conducive to provincial economic development, it is only necessary to verify whether PFTZ implementation affects these mediating variables. As shown in

Table 9. Mediating effect tests.

	(1)	(2)	(3)	(4)
	LnAGDP	LnFDI	Apply	Market
did	0.075 ***	0.256 ***	3.006 ***	−0.014
	(2.738)	(3.152)	(3.522)	(−0.171)
UL	0.036 ***	0.012	0.944 ***	0.165 ***
	(5.840)	(0.675)	(4.928)	(9.231)
EL	−7.459	47.516 ***	−734.554 ***	−55.699 ***
	(−1.225)	(2.625)	(−3.858)	(−3.129)
IL	0.000	−0.000	−0.001	0.000
	(0.070)	(−0.392)	(−0.352)	(0.330)
FTD	−0.338 **	0.306	−42.475 ***	−1.189 ***
	(−2.314)	(0.707)	(−9.308)	(−2.786)
FE	−0.672 **	0.113	39.934 ***	2.293 **
	(−1.971)	(0.112)	(3.748)	(2.302)
_cons	8.790 ***	11.038 ***	−26.888 ***	1.113
	(33.607)	(14.202)	(−3.289)	(1.456)
N	420	420	420	420
R2	0.922	0.923	0.859	0.958
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, the estimated coefficient of the dummy variable (did) on capital flow (LnFDI) is positive, indicating that, compared with the control group, the actual use of FDI in the experimental group is 25.6% higher. This demonstrates that PFTZ implementation effectively facilitates capital inflow.

Regarding technological innovation (apply), the regression coefficient of the core explanatory variable is 3.006 and significant at the 1% level, indicating that PFTZ implementation significantly enhances provincial innovation capacity. However, the result for market-oriented reforms (market) is statistically insignificant, indicating that the driving force of PFTZs for market reform has not yet materialized. On the one hand, this may be due to the close association between PFTZ implementation and administrative streamlining reforms, which could affect the redistribution of vested interests. The loss of certain powers may slow the reform progress and delay its effects [79,80]. On the other hand, this may also be related to differences in the content of market-oriented reforms across various zones. For example, initiatives such as the “22-in-1 permit” and “48-certificate joint processing” systems implemented in the Kaifeng Pilot Zone in recent years, while having certain benefits, may create challenges for cross-zone operations due to differing approval authorities across zones [29].

4.6. Further Analysis

4.6.1. Regional Mediating Effect

As shown in

Table 10. Eastern and central-western regions-Capital-mediating effect test.

	(1)	(2)	(3)	(4)
	Original Form	Capital-Mediating Effect Test	Central-Western Region	Eastern Region
	LnAGDP	LnFDI	LnFDI	LnFDI
did	0.075 ***	0.256 ***	0.185 **	0.307 **
	(2.738)	(3.152)	(2.122)	(2.159)
UL	0.036 ***	0.012	0.105 ***	−0.105 ***
	(5.840)	(0.675)	(2.725)	(−2.958)
EL	−7.459	47.516 ***	68.996 ***	107.054 ***
	(−1.225)	(2.625)	(2.656)	(2.645)
IL	0.000	−0.000	−0.000 *	0.155 **
	(0.070)	(−0.392)	(−1.715)	(2.364)
FTD	−0.338 **	0.306	0.792	1.986 ***
	(−2.314)	(0.707)	(0.771)	(2.887)
FE	−0.672 **	0.113	−0.697	6.003 *
	(−1.971)	(0.112)	(−0.671)	(1.928)
_cons	8.790 ***	11.038 ***	6.831 ***	16.665 ***
	(33.607)	(14.202)	(4.382)	(9.823)
N	420	420	238	140
R2	0.922	0.923	0.929	0.815
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, PFTZ implementation has a statistically significant impact on capital flow in eastern and central-western regions. Notably, the estimated coefficient for the eastern region (0.307) is substantially larger than that for the central-western regions (0.185), indicating that PFTZs in eastern China have a stronger effect in promoting capital flow. This regional disparity may be attributed to the east’s concentration of advanced coastal cities and the region’s comparatively higher degree of openness to international markets. These structural advantages enhance the eastern region’s ability to attract actual use FDI, subsequently strengthening the capital flow effects of PFTZ implementation.

Regional heterogeneity is also reflected in the performance of the control variables, which reinforces the baseline conclusions. In the eastern region, nearly all control variables are statistically significant and predominantly positive, whereas in the central and western regions, several variables (e.g., FTD and FE) are insignificant (shown in Table 10). In particular, the influence of education on attracting FDI appears much stronger in the eastern (107.054) than in central and western regions (68.996), highlighting the gradient differences in the contribution of human capital. Similarly, IL, FTD, and FE only demonstrate significant positive effects on FDI in the eastern region (IL: 0.155; FTD: 1.986; FE: 6.003). High R² values (ranging from 0.815 to 0.929) and the application of a double FE model for the full sample and subsamples further support the robustness of these findings, collectively demonstrating that the capital mediation mechanism operates with significantly greater effectiveness in the eastern region.

The results in column (2) of

Table 11. Eastern and central-western regions-Technology-mediating effect test.

	(1)	(2)	(3)	(4)
	Original Form	Technology-Mediating Effect Test	Central-Western Region	Eastern Region
	LnAGDP	Apply	Apply	Apply
did	0.075 ***	3.006 ***	2.478 ***	−2.424
	(2.738)	(3.522)	(5.925)	(−1.262)
UL	0.036 ***	0.944 ***	0.230 *	2.792 ***
	(5.840)	(4.928)	(1.845)	(7.176)
EL	−7.459	−734.554 ***	−175.305 **	−1002.904 **
	(−1.225)	(−3.858)	(−2.075)	(−2.547)
IL	0.000	−0.001	0.000	−1.437 ***
	(0.070)	(−0.352)	(0.671)	(−4.469)
FTD	−0.338 **	−42.475 ***	14.011 ***	−74.059 ***
	(−2.314)	(−9.308)	(4.265)	(−6.070)
FE	−0.672 **	39.934 ***	23.121 ***	36.446
	(−1.971)	(3.748)	(5.369)	(1.156)
_cons	8.790 ***	−26.888 ***	−8.064	−162.943 ***
	(33.607)	(−3.289)	(−1.496)	(−5.198)
N	420	420	238	140
R2	0.922	0.859	0.868	0.898
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

examine the mediating role of apply, and columns (3) and (4) present the results of the mediating effect tests after stratifying the sample into central-western and eastern regions. The findings reveal that PFTZ implementation has a statistically significant positive impact on the central-western regions’ technological innovation, with a coefficient of 2.478 (p < 0.01). This regional difference may be explained by two factors. First, the eastern region has a higher baseline level of patent grants compared with the central-western region, leaving more room for technological advancement in the latter. Second, although newly established high-tech enterprises in China remain concentrated in the eastern and southern regions, recent policy incentives have promoted the diffusion of such enterprises into central and western areas. Accordingly, the observed mediating effects reflect regional development disparities and the redistribution of innovation driven by national policy.

4.6.2. Economic Scale Mediating Effect

As shown in

Table 12. Size of the economy-Capital-mediating effect test.

	(1)	(2)	(3)	(4)
	Original Form	Capital-Mediating Effect Test	Lower Level of Economic Development	Higher Level of Economic Development
	LnAGDP	LnFDI	LnFDI	LnFDI
did	0.075 ***	0.256 ***	0.582 ***	0.092
	(2.738)	(3.152)	(3.742)	(1.482)
UL	0.036 ***	0.012	−0.094 *	0.017
	(5.840)	(0.675)	(−1.749)	(1.127)
EL	−7.459	47.516 ***	62.998 *	7.709
	(−1.225)	(2.625)	(1.811)	(0.553)
IL	0.000	−0.000	0.000	−0.035 **
	(0.070)	(−0.392)	(0.709)	(−2.344)
FTD	−0.338 **	0.306	−5.208 ***	0.600 **
	(−2.314)	(0.707)	(−2.819)	(2.217)
FE	−0.672 **	0.113	−3.595 ***	4.195 ***
	(−1.971)	(0.112)	(−2.619)	(3.725)
_cons	8.790 ***	11.038 ***	16.769 ***	11.380 ***
	(33.607)	(14.202)	(6.324)	(10.862)
N	420	420	210	210
R2	0.922	0.923	0.885	0.965
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, the regression coefficient for less-developed regions is notably larger than that for more-developed regions, with the former exhibiting a positive and statistically significant effect at the 1% level. This indicates that capital flow functions as a more effective channel through which PFTZ implementation generates economic impact in regions with lower economic development. This may be attributed to the fact that regions with higher economic development are inherently better positioned to attract and retain a large number of high-quality enterprises, capital, and talent, among other key factors. These factors lay a solid foundation for the region’s economic growth. Therefore, when such regions receive policy support, it may paradoxically weaken the marginal effects of Pilot Free Trade Zones. For regions with lower economic development, which have relatively fewer high-quality enterprises, capital, and talent, there is significant room for improvement. The implementation of PFTZ can further enhance infrastructure development, talent attraction, high-end industrial development, and technological innovation in these regions, thereby helping them overcome key thresholds in economic development. In all, PFTZs play a more effective role in promoting economic growth in regions with lower economic development [81]. It also implies that capital mobility is a critical mechanism for less-developed regions to strengthen their economic capacity, mitigating structural disadvantages by enhancing the efficiency of policy-driven capital allocation.

As demonstrated in

Table 13. Size of the economy-Technology-Mediating effect test.

	(1)	(2)	(3)	(4)
	Original Form	Technology-Mediating Effect Test	Higher Level of Economic Development	Lower Level of Economic Development
	LnAGDP	Apply	Apply	Apply
did	0.075 ***	3.006 ***	−0.116	2.112 ***
	(2.738)	(3.522)	(−0.080)	(5.358)
UL	0.036 ***	0.944 ***	2.169 ***	0.731 ***
	(5.840)	(4.928)	(6.454)	(7.481)
EL	−7.459	−734.554 ***	−886.746 ***	−235.569 ***
	(−1.225)	(−3.858)	(−3.677)	(−2.777)
IL	0.000	−0.001	−1.047 ***	−0.001 *
	(0.070)	(−0.352)	(−4.715)	(−1.669)
FTD	−0.338 **	−42.475 ***	−53.142 ***	22.265 ***
	(−2.314)	(−9.308)	(−5.523)	(5.871)
FE	−0.672 **	39.934 ***	88.332 ***	19.340 ***
	(−1.971)	(3.748)	(3.804)	(5.807)
_cons	8.790 ***	−26.888 ***	−136.593 ***	−29.208 ***
	(33.607)	(−3.289)	(−5.429)	(−6.889)
N	420	420	210	210
R2	0.922	0.859	0.889	0.897
province effect	Yes	Yes	Yes	Yes
time effect	Yes	Yes	Yes	Yes

Note: *, **, and *** indicate significant at the 10, 5, and 1 percent levels, respectively.

, the regression coefficient for less-developed regions is positive and statistically significant (2.112, p < 0.01), whereas the result for more-developed regions is statistically insignificant. This indicates that PFTZ implementation has no notable impact on patent grants in economically advanced regions, indicating a weak correlation between PFTZs and technological innovation outcomes in these areas. In contrast, in less-developed regions, PFTZs significantly enhance technological innovation capacity. By translating these innovation capabilities into patent development, less-developed regions can leverage policy-driven mechanisms to promote economic growth more effectively, contributing to reducing regional development disparities through institutional innovation.

Empirical results confirm that technological innovation mechanism is a mediating mechanism for promoting sustainable economic development in PFTZs, and that the effectiveness of this mechanism varies significantly across regions with different levels of economic development. This means that PFTZs can help overcome regional disparities and promote technological innovation catch-up through a series of institutional innovations.

5. Conclusions

Based on the empirical findings and relevant literature, this chapter summarizes the key findings and policy implications regarding economic growth effects, transmission mechanisms, and heterogeneity analysis. It also outlines the study’s limitations and outlines potential directions for future research.

5.1. Summary and Discussion

Using provincial panel data spanning 2010 to 2023, this study examines a sample of 30 provincial-level administrative regions in China and applies a multiperiod DID approach to systematically examine the impact mechanisms and heterogeneity of PFTZ implementation on provincial economic growth. The empirical findings yield three key conclusions. (1) After controlling for the potential overlapping policy effects of the BRI, a series of robustness tests, including parallel trend and placebo tests, confirm that PFTZ implementation significantly promotes regional economic growth. The estimated policy effect remains statistically significant after accounting for exogenous shocks (p < 0.01), indicating that PFTZs are a fundamental driver of provincial economic development. (2) Mediating effect analyses reveal that PFTZs primarily stimulate economic growth through enhanced capital flow and improved technological innovation transmission mechanisms. However, the moderating role of market-oriented reforms is not statistically significant (p > 0.1), potentially due to the inherent redistribution of benefits and institutional resistance embedded in such reforms. (3) Heterogeneity analysis indicates that PFTZ implementation has a more pronounced effect in central and western provinces and economically underdeveloped regions. This disparity may be attributed to the greater scope for institutional innovation in late-developing areas and the diminishing marginal returns associated with policy interventions.

This study systematically examines the driving mechanisms and heterogeneity of PFTZs in promoting sustainable regional economic growth. The empirical results confirm that PFTZ policy significantly enhances provincial economic sustainability, and this effect remains robust even after controlling for the overlapping influence of the BRI policy, underscoring the core value of PFTZs as testing grounds for institutional innovation. These findings are consistent with the conclusions of other scholars in this field [37].

It is worth noting that most existing studies have focused on indicators such as import and export volumes or trade structure [16,82]. While they also find that PFTZs promote economic growth, this study takes a different perspective by focusing on per capita GDP as the core proxy for regional economic sustainability, thereby enriching the research dimensions on the economic effects of PFTZs. As an important national strategy, PFTZs have become a key driver for strengthening enterprise innovation, developing characteristic industries, and fostering regional economic growth [83,84]. Yet, their underlying mechanisms remain underexplored. To address this gap, this study integrates the neoclassical growth model, endogenous growth theory, and relevant literature [85,86] to construct a “capital–technology–marketization” framework, revealing the pathways through which PFTZs drive sustainable development. The mechanism analysis shows that PFTZs primarily stimulate growth by enhancing capital liquidity and strengthening the transmission of technological innovation. These findings align with prior studies [70,87], confirming the synergistic effects of capital and technology. However, the marketization channel did not show significant mediating effects, a result for which we provide explanatory evidence, offering new insights into this field.

By employing provincial-level data, this study bridges the gap between macro- and micro-level research on PFTZs and validates the mechanisms through which they contribute to economic growth. Future research should further investigate the interactions among capital, technology, and marketization; explore ways to address the barriers to marketization reforms in less-developed regions; and examine the long-term dynamic impacts of PFTZ policy using micro-level enterprise panel data.

5.2. Theoretical Contributions

First, this study employs a multiperiod DID model to disentangle the overlapping effects of the PFTZ policy and regional strategies such as the BRI more accurately and clarify the independent driving influence of PFTZ implementation in promoting provincial economic growth. Second, the study constructs a three-dimensional transmission mechanism framework encompassing capital–technology–marketization, revealing the core mediating influences through which PFTZs exert economic effects, and filling a gap in systematic empirical research and multidimensional mechanism analysis in this field. Third, the study identifies deeper causes behind development disparities across heterogeneous regions and areas with varying levels of economic development, providing a new micro-foundation for promoting coordinated and sustainable regional development.

5.3. Policy Recommendations

The empirical findings clearly demonstrate that the establishment of PFTZs has a positive impact on regional economic sustainability. However, achieving high-quality, long-term development also requires a solid ecological foundation, as economic prosperity ultimately depends on the carrying capacity of natural resources and the environment. Therefore, the next stage of PFTZ policy design should place greater emphasis on aligning economic growth objectives with ecological and environmental protection. Ensuring that growth is not only rapid, but also green, inclusive, and sustainable, has become a central policy imperative.

Building on the empirical results, this study proposes several policy recommendations. First, institutional innovation within PFTZs should be further deepened to foster high-quality and sustainable development. The PFTZ strategy should advance under a dual-engine approach that combines replication of successful experiences with context-specific innovation. Newly established PFTZs should integrate proven practices from earlier pilots while adapting development models to local resource endowments and stages of development—particularly in areas such as green transformation, low-carbon growth, and ecological conservation. Policymakers should also prioritize the establishment of PFTZ subzones in provincial capitals of central and western regions, as well as in strategically important border areas, to cultivate new regional growth poles with strong environmental resilience supported by institutional transfer and policy coordination.

In parallel, a nationwide PFTZ collaboration network should be established to promote cross-regional cooperation in innovation, industrial integration, and policy diffusion. Such a network would enhance knowledge sharing, strengthen industrial linkages, and ensure that regions at different development levels can all benefit from long-term, stable, and green economic growth. In this way, the PFTZ system can function not only as a platform for economic liberalization, but also as a driver of China’s broader goals of green and balanced regional development.

Secondly, strengthening the institutional mechanisms that underpin PFTZ effectiveness is essential for ensuring sustainable economic growth. Since capital mobility is a key channel through which PFTZs drive regional development, a well-designed two-way capital management system should be established to channel financial flows into sustainable sectors. This entails streamlining cross-border investment and financing procedures, enhancing full-cycle services for foreign investors, and prioritizing support for green investment projects.

Moreover, policymakers should strengthen the strategic focus on technological innovation—particularly in low-carbon and environmentally sustainable fields—by fostering an integrated innovation ecosystem that links basic research, applied development, and commercial deployment. To accelerate this process, targeted incentive policies such as R&D subsidies and support for equipment procurement should be directed toward firms engaged in breakthrough or bottleneck technologies, especially in green industries. These measures will enhance the capacity of PFTZs to achieve balanced progress in both economic growth and environmental sustainability [88].

5.4. Research Limitations

This study’s exploration of the pathways through which PFTZ implementation promotes provincial economic growth leaves room for further refinement. First, regarding the depth of mechanism analysis, while the key pathways of capital, technology, and marketization have been identified, further qualitative and empirical research is needed to examine their interactions and the micro-level dynamics within each pathway. In particular, the results show that marketization has not yet played a significant mediating role, a phenomenon that warrants deeper investigation using firm- or household-level data. Second, with respect to institutional heterogeneity, PFTZs established in different phases, regions, and strategic contexts exhibit varying priorities in institutional innovation and openness. Although this study provides differentiated policy recommendations, empirical analysis has not fully deconstructed how such heterogeneity shapes economic outcomes across different PFTZ stages. Third, in terms of research scope, this study has primarily focused on economic growth. While GTFP was incorporated to account for environmental factors, future work should more fully integrate micro-level indicators (e.g., corporate performance), environmental dimensions (e.g., carbon emissions, resource utilization, green technology adoption), and social outcomes (e.g., employment, inequality, urban–rural integration). Accordingly, future research should broaden its focus to capture the comprehensive impacts of PFTZs on institutional innovation, regulatory alignment, government transformation, social governance, and green development, as well as their interactions with economic outcomes. Finally, because this study relies on provincial-level data, it faces limitations in scope. Collecting prefecture-level data and extending the dataset to more recent years will be essential for deepening the analysis of PFTZ policy effects.