FDI Spillovers and High-Quality Development of Enterprises—Evidence from Chinese Service Enterprises

Abstract

The core of high-quality development lies in achieving long-term sustainability. In the context of China’s high-quality economic development and high-standard opening-up of the service industry, it is of great theoretical value and practical significance to study how service enterprises can effectively absorb foreign direct investment (FDI) spillovers to realize high-quality development and enhance sustainable value creation capability. Based on the panel data of A-share non-financial listed service enterprises in China, this study explores the impact of FDI on the high-quality development of service enterprises in China through various spillover channels, as well as the role of enterprise absorptive capacity in the relationship between FDI and high-quality development. The results show that horizontal and backward spillovers have a significant positive impact on the high-quality development of service enterprises, while forward spillovers have a significant negative impact. Heterogeneity analysis indicates that the promotion effect of horizontal spillovers is more pronounced on enterprises in the eastern region, capital-intensive enterprises, small and medium-sized enterprises (SMEs), and producer service enterprises. The promotion effect of backward spillovers is particularly evident for enterprises in the central and western regions, capital-intensive enterprises, SMEs, non-state-owned enterprises (non-SOEs), and producer service enterprises. The further threshold regression model finds that service enterprises with higher absorptive capacity benefit more through horizontal and vertical spillovers for their high-quality development.

1. Introduction

High-quality development is an “upgraded version” of sustainable development [1], signifying that China’s economy is embarking on a path of healthy and sustainable growth [2]. Currently, China is in the era of a service economy. High-quality development of the service industry is of great significance in optimizing industrial structure, creating employment opportunities, enhancing national competitiveness, and achieving environmental protection and sustainable development. As the micro-foundations of macroeconomics, the high-quality development of enterprises is the basis and focus of realizing the high-quality development of industries and the economy. The goal of high-quality development of enterprises is to pursue sustained growth and value creation capacity [3], including the comprehensive benefits of innovative development, green development, and sustainable development.

High-quality development is inseparable from high-standard opening-up. In recent years, China has been confronted with the significant tasks of optimizing and upgrading the industrial structure, adjusting the economic development mode, and establishing an international and domestic dual circulation development pattern. Foreign investment, as a bond connecting domestic and international dual circulations, plays a significant role in constructing the new development paradigm and facilitating the virtuous cycle and high-quality development of the national economy. In 2021, the proportion of foreign investment utilized in China’s service industry was as high as 76.3%, reaching a historical peak and achieving growth against the trend of the global economic downturn. From 2022 to 2023, the proportion of foreign investment utilized in China’s service industry has declined, but the high-tech service industry, which represents the trend of industrial upgrading, has been outstanding in attracting foreign investment. The Chinese government has proposed that greater emphasis should be placed on enhancing the quality of foreign investment utilization and increasing support for foreign investment in the modern service industry. Therefore, investigating the mechanisms through which Chinese service enterprises can effectively absorb foreign direct investment (FDI) spillovers to attain high-quality development holds substantial theoretical and practical significance.

Since the concept of “high-quality development” was first proposed in China, scholars have interpreted its connotation at the macro, meso, and micro levels. Huang et al. [3] put forward that in contrast to the unsustainable development pattern, the high-quality development of enterprises is an efficient, equitable, green, and sustainable development with high development quality. Research on how to measure it is usually divided into two categories: single-indicator evaluation and multi-indicator comprehensive evaluation. The single indicator evaluation method mainly measures high-quality development through total factor productivity (TFP) [4,5]. The multi-indicator comprehensive evaluation mostly employs principal component analysis and the entropy weight method to construct an index system. Liao et al. [6] constructed an index system from the perspectives of innovation level, TFP, enterprise growth, production and operation efficiency, and enterprise scale. Li et al. [7] constructed an indicator system covering three dimensions of “economic-social-environmental”.

A number of empirical studies have examined the FDI spillover effects of Chinese manufacturing enterprises through aspects of enterprise productivity, export performance, innovation, wage levels, and enterprise survival. Several studies have confirmed the existence of positive FDI spillover effects through horizontal or vertical linkages. Wei and Liu [8] found positive intra- and inter-industry productivity spillovers from FDI to Chinese manufacturing firms within regions. Lin et al. [9] found significant vertical spillover effects of FDI on Chinese domestic manufacturing firms. Xu and Sheng [10] suggested that positive spillovers from FDI arise from forward linkages. Li et al. [11] pointed out that high-tech manufacturing firms can improve their productivity through their connections with foreign-invested firm customers. While most existing studies on China argued that FDI has positive vertical spillover effects, the existing literature has not reached a consensus on the horizontal effects of FDI. Lu et al. [12] found that FDI has a negative impact on the productivity of domestic enterprises in China’s manufacturing sector through horizontal spillovers. Other studies also suggested that horizontal spillovers are influenced by the origin of FDI and the characteristics of different industries. Lin et al. [9] found that FDI from Hong Kong, Macao, and Taiwan (HMT) generates negative horizontal spillovers, while FDI from non-HMT regions tends to generate positive horizontal spillovers in China. Jeon et al. [13] demonstrated that the negative effects of horizontal spillovers are particularly pronounced in low-technology sectors. Despite the growing literature concerning FDI spillovers, prior research has rarely investigated the effects of FDI spillovers on service enterprises.

Research on service FDI spillovers is mostly conducted at national, regional, and industrial levels, and the findings remain inconclusive. Most studies have suggested that FDI in the service industry has a positive spillover effect. Doytch and Uctum [14] conducted a study based on data from 14 service sectors in 14 APEC economies, finding that FDI in the service sector has a significant positive spillover effect on GDP growth in general. Xu and Liu [15] indicated that FDI has a positive spillover effect on the independent innovation of the service sector. However, there are also a few studies suggesting the insignificant and even negative FDI spillover effects in the service industry. Ingham et al. [16] conducted a sector-specific study on Egypt and found that FDI in some service sectors has significant negative spillover effects. Previous studies on service FDI spillovers have rarely examined the influence of different types of FDI spillovers. Ignoring the channels through which spillover effects arise may overestimate the influence of FDI on domestic enterprises [17]. Although scholars have already identified absorptive capacity as an important internal factor influencing the spillover effects of FDI, the common practice is to measure the absorptive capacity of enterprises using a single indicator such as human capital [18], R&D investment [19], and technological gap [20]. Paola and Rajneesh [21] argued that absorptive capacity cannot be assessed from a single perspective. Instead, the “synergistic effects” of different influential factors on absorptive capacity should be taken into consideration. Therefore, it is highly necessary to select the determinants of absorptive capacity and construct the corresponding quantitative index system in the empirical examination of absorptive capacity.

Currently, research on the factors affecting the high-quality development of enterprises primarily explores two perspectives: external factors and internal factors. External factors such as the digital economy [22], digital finance [4], and high-speed railway [7] help to promote the high-quality development of enterprises, while climate change inhibits the high-quality development of enterprises [5]. There is a U-shaped relationship between environmental regulation and enterprise high-quality development [23]. Internal factors such as digital transformation [24], digitalization and green transformation [25], ESG performance [26], entrepreneurs [27], and corporate social responsibility [28] help enterprises achieve high-quality development, while corporate risk-taking has a significant negative impact [29]. Some studies have shown the positive impact of FDI on high-quality economic development [2,30]. However, there is a scarcity of studies on the relationship between FDI and the high-quality development of enterprises, especially service enterprises.

The major contributions of this study can be described as follows. First, this study supplements the existing research by enhancing the understanding of FDI spillover effects in the service industry of developing economies. Second, this study clarifies the impact of different spillovers on the high-quality development of service enterprises. Third, this study considers the comprehensive and dynamic nature of absorptive capacity and constructs an absorptive capacity measurement system, which enriches the study of the absorptive capability indicator system. Finally, this study provides new micro-level evidence supporting the existence of absorptive capacity thresholds in different types of service FDI spillovers. The findings of this study contribute to the formulation of effective policies to utilize FDI spillovers for the high-quality development of service enterprises.

2. Theoretical Analysis and Research Hypothesis

2.1. FDI Spillovers and the High-Quality Development of Service Enterprises

Currently, there is a scholarly consensus regarding the channels through which FDI spillovers are generated. The entry of multinational corporations (MNCs) into a host country brings capital, technology, and human resources, positioning local firms as competitors, suppliers, or customers of the MNCs [18]. Consequently, FDI influences the high-quality development of local enterprises via three distinct spillover channels, namely, horizontal, forward, and backward spillovers. The impact of FDI results from the synergistic effect of these three spillover channels.

From the perspective of the horizontal spillovers, first, it is essential to note that, given the distinctiveness of the service industry, in addition to “hard technologies” such as equipment and production processes, “soft technologies” like management experience and marketing practices are critically important for service enterprises [31]. When communicating with foreign enterprises, domestic enterprises can enhance their knowledge accumulation and innovation through the imitation of foreign enterprises. Consequently, they can effectively attain positive demonstration and imitation effects regarding “soft technologies”, thereby improving their TFP to achieve high-quality development. Second, enterprises in some important service sectors in China are mostly state-owned enterprises with serious administrative monopolies. Market-seeking-oriented foreign investment in the service sector can generate a competitive effect and reduce market monopoly. On the one hand, competition creates positive incentives. To ensure market share, local service enterprises will attach importance to innovation, increase R&D investment, and enhance the service level, thus realizing high-quality development. On the other hand, competition also produces a “crowding-out effect”. Enterprises that lack competitiveness will be eliminated by the market, and resources will flow to high-efficiency enterprises, realizing the reintegration of local enterprises’ resources. Third, management knowledge in the service industry is often internalized in people’s thinking patterns, so human capital plays an important role in knowledge spillover and diffusion. Foreign service enterprises need to rely on the human capital of the host country for their production and operation in the local market. Labor mobility between MNCs and local enterprises is an important channel for knowledge dissemination in the service sector. Meanwhile, the good human resource management experience of MNCs will also be imitated by domestic service enterprises, which lays an important foundation for domestic enterprises to accumulate human resource experience for their innovative activities.

From the perspective of vertical spillover effects, compared to manufacturing FDI, the “localization” characteristics of service FDI result in stronger vertical spillover effects. To better grasp the consumption demands and customer preferences, as well as to provide targeted services to customers in the host country market, foreign service enterprises have a greater need to establish close supply-and-demand connections with local service enterprises [32].

Theoretically, the forward spillover effect can be either positive or negative. On the one hand, the diffusion effect of the service industry is relatively strong. Upstream enterprises can more easily transfer knowledge to downstream enterprises, thereby influencing the production decisions and business performance of downstream enterprises [33]. Meanwhile, the expansion of foreign investment in the service industry can introduce competition into the domestic market for intermediate service items, facilitating the reduction in prices and improvement in the quality of domestic service offerings. The high-quality and low-cost service inputs directly lower the production costs of downstream enterprises, alleviate financing constraints, and create an enabling environment for productivity growth. On the other hand, MNCs can employ monopolistic pricing with price markups for differentiated products, and high-quality inputs might imply higher prices. If domestic enterprises fail to gain sufficiently large benefits, negative spillover effects will arise.

Some scholars argued that backward spillover is the channel most likely to generate positive spillovers, as MNCs are strongly motivated to share knowledge with their suppliers [34]. Firstly, to obtain high-quality inputs, downstream foreign enterprises will collaborate with domestic suppliers on technological innovation, personnel training, and management training, thereby generating a technology diffusion effect. Secondly, even domestic enterprises that have no cooperative relationship with foreign enterprises may benefit from the scale economies resulting from increased quality and diversification in the intermediate goods market. Given the analysis above, this study proposes the following hypotheses.

Hypothesis 1a. 

FDI can promote the high-quality development of service enterprises through horizontal spillovers.

Hypothesis 1b. 

FDI can promote the high-quality development of service enterprises through backward spillovers.

Hypothesis 1c. 

FDI can promote the high-quality development of service enterprises through forward spillovers.

2.2. Threshold Effect of Enterprise Absorptive Capacity

The absorptive capacity of enterprises serves as a crucial internal determinant that shapes the nexus between FDI spillovers and the high-quality development of service enterprises. In emerging economies where significant gaps exist between foreign and domestic enterprises, the absorptive capacity becomes even more crucial [35]. A positive TFP effect derives from the interaction and complementarity effect of the ability to acquire, assimilate, transform, and exploit external knowledge [36].

On the one hand, from the perspective of horizontal spillover effects, domestic service enterprises with higher knowledge acquisition ability can imitate and learn valuable external knowledge through increasing communication and collaboration with external partners [36]. The ability to digest and integrate knowledge can maximize the spillover effects of FDI and enable enterprises to achieve independent innovation. On the other hand, under competitive pressure from foreign enterprises, domestic enterprises with high absorptive capacity will intensify their R&D investment and innovation initiatives to cope with the competition. The study by Han et al. [37] found that FDI exerts a positive competitive effect on enterprises with high absorptive capacity, indicating that the stronger the absorptive capacity, the stronger the competitiveness of the enterprise. Enhanced absorptive capacity in domestic firms compels MNCs to intensify technological innovation and forge new competitive edges. These evolving MNCs’ advantages create opportunities for domestic enterprises to capture enhanced spillover effects. The enhancement of absorptive capacity amplifies the demonstration–imitation and competition effects at advanced stages, thereby upgrading the quality of FDI spillovers to facilitate sustained high-quality development in domestic enterprises.

From the perspective of vertical spillover effects, the absorptive capacity of domestic enterprises will first influence the decision-making of MNCs regarding the selection of suppliers. Duong and Vinh [38] discovered that domestic Vietnamese firms with good absorptive capacity have a higher chance of being suppliers to Japanese firms. Upstream local suppliers with a strong industrial chain and good absorptive capacity can meet the breadth and depth of knowledge required by MNCs in extensive technological collaboration and knowledge and information sharing. Host country suppliers with strong absorptive capacity can also integrate the requirements of MNCs for suppliers with their strategic development, thus enhancing their independent innovation capacity. Second, downstream domestic firms need sufficient absorptive capacity to utilize the high-quality inputs provided by upstream MNCs effectively, thereby improving productivity and achieving technological innovation [18]. If an enterprise lacks the capacity for absorption and independent innovation and excessively relies on inputs provided by MNCs, it will fall into the vicious cycle of continuous introduction and falling behind, thereby losing the potential for sustained growth and value creation.

From the above analysis, it can be concluded that absorptive capacity can moderate the relationship between FDI spillovers and the high-quality development of service enterprises. An enhanced absorptive capacity of enterprises is conducive to their more efficient utilization of FDI spillovers and ultimately facilitates the high-quality development of enterprises. The impact of absorptive capacity on the linkage between FDI and domestic enterprises may be nonlinear [39]. Several studies have shown the threshold effect of absorptive capacity in FDI spillover effects [40,41,42]. These studies have suggested that firms need to possess a certain level of absorptive capacity to successfully exploit spillover effects. On this basis, the second hypothesis of this study is proposed.

Hypothesis 2. 

There is a threshold effect of absorptive capacity in the impact of FDI spillovers on the high-quality development of service enterprises.

According to the above analysis, the theoretical analysis framework is presented in Figure 1.

3. Study Design

3.1. Model Construction

Following the earlier work by Jude [34], Orlic et al. [43], and Ha et al. [18], to test the direct impact of FDI spillovers on the high-quality development of China’s service enterprises, a benchmark model was constructed as follows:

$$TF{P}_{ijt}={\beta }_0+{\beta }_{1}h{s}_{jt}+{\beta }_{2}f{s}_{jt}+{\beta }_{3}b{s}_{jt}+\sum \gamma fir{m}_{it}+hh{i}_{jt}+{\delta }_i+{\sigma }_t+{\epsilon }_{ijt}$$

(1)

In Equation (1), the subscripts i, j, and t represent enterprise, industry, and year, respectively. $TF{P}_{ijt}$ denotes the high-quality development of service enterprises. $h{s}_{jt}$, $f{s}_{jt}$, and $b{s}_{jt}$ denote horizontal, forward, and backward spillovers, respectively. $\sum \gamma fir{m}_{it}$ denotes a set of control variables at the firm level. $hh{i}_{jt}$ denotes the degree of market competition at the industry level and is an industry-level control variable. ${\delta }_i$ denotes the individual fixed effects. ${\sigma }_t$ denotes the time-fixed effects and ${\epsilon }_{ijt}$ is the random error term.

When estimating moderating effects, two common approaches are the grouped regression method and the interaction term method. However, the grouped regression results are sensitive to subjective grouping criteria, while the interaction term approach may suffer from multicollinearity issues. Crucially, neither method can accurately estimate the thresholds of moderating variables. Hansen’s [44] threshold model cannot only estimate threshold values but also test their accuracy through statistical significance tests. We thus use the threshold regression model introduced by Hansen [44] to identify the threshold of absorptive capacity and compare the FDI productivity nexus when the strength of absorptive capacity is below and above this threshold. Based on Equation (1), a multiple threshold panel regression model with enterprise absorptive capacity as the threshold variable (taking double threshold as an example) is constructed.

$$\begin{array}{l}TF{P}_{ijt}={\beta }_0+{\beta }_{1}f{s}_{jt}+{\beta }_{2}b{s}_{jt}+{\delta }_{1}h{s}_{jt}I(absor{b}_{ijt}<{\theta }_1)+{\delta }_{2}h{s}_{jt}I({\theta }_1\le absor{b}_{ijt}<{\theta }_2)\\ +{\delta }_{3}h{s}_{jt}I(absor{b}_{ijt}\ge {\theta }_2)+\sum \gamma fir{m}_{it}+hh{i}_{jt}+{\delta }_i+{\sigma }_t+{\epsilon }_{ijt}\end{array}$$

(2)

$$\begin{array}{l}TF{P}_{ijt}={\beta }_0+{\beta }_{1}h{s}_{jt}+{\beta }_{2}b{s}_{jt}+{\delta }_{1}f{s}_{jt}I(absor{b}_{ijt}<{\theta }_1)+{\delta }_{2}f{s}_{jt}I({\theta }_1\le absor{b}_{ijt}<{\theta }_2)\\ +{\delta }_{3}f{s}_{jt}I(absor{b}_{ijt}\ge {\theta }_2)+\sum \gamma fir{m}_{it}+hh{i}_{jt}+{\delta }_i+{\sigma }_t+{\epsilon }_{ijt}\end{array}$$

(3)

$$\begin{array}{l}TF{P}_{ijt}={\beta }_0+{\beta }_{1}h{s}_{jt}+{\beta }_{2}f{s}_{jt}+{\delta }_{1}b{s}_{jt}I(absor{b}_{ijt}<{\theta }_1)+{\delta }_{2}b{s}_{jt}I({\theta }_1\le absor{b}_{ijt}<{\theta }_2)\\ +{\delta }_{3}b{s}_{jt}I(absor{b}_{ijt}\ge {\theta }_2)+\sum \gamma fir{m}_{it}+hh{i}_{jt}+{\delta }_i+{\sigma }_t+{\epsilon }_{ijt}\end{array}$$

(4)

where $absor{b}_{ijt}$ denotes the threshold variable, which is the absorptive capacity of the enterprise. I(*) is an indicator function that takes the value of 1 when the conditions within the parentheses are met and 0 otherwise. ${\theta }_1$ and ${\theta }_2$ are the estimates of the dual-threshold. ${\delta }_1$, ${\delta }_2$, and ${\delta }_3$ denote the corresponding coefficients of the explanatory variables at different thresholds. Since the threshold panel regression model cannot set multiple core explanatory variables simultaneously, this study takes the horizontal spillover effect ($h{s}_{jt}$), the forward spillover effect ($f{s}_{jt}$), and the backward spillover effect ($b{s}_{jt}$) as the core explanatory variables, respectively, to measure the threshold values of absorptive capacity in each spillover effect.

3.2. Selection of Variables

3.2.1. Explained Variable

High-quality development of enterprises (TFP): TFP is an indicator of output growth that comprehensively measures enterprise efficiency and reflects the development level of an enterprise [4]. Therefore, this study employs TFP as the explained variable to measure the high-quality development of enterprises. The LP method proposed by Levinsohn and Petrin [45] is utilized in this study to estimate TFP, and in the robustness test, the ACF method is used to calculate TFP.

3.2.2. Explanatory Variables

• Horizontal spillovers (hs)

Sales is a common indicator utilized in the current literature to measure horizontal spillovers [46]. Given that the research subjects of this study are listed service companies, based on data availability and referring to the studies of Jude [34] and Orlic et al. [43], with the foreign equity ratio of enterprises as the weight, this study calculates the proportion of the main business revenue of MNCs to the total main business revenue of the industry to measure the horizontal spillover effect:

$$h{s}_{jt}={\displaystyle \frac{{\displaystyle \sum _{i\in {\theta }_j}{\rho }_{ijt}\times y_{ijt}}}{{\displaystyle \sum _{i\in {\theta }_j}{y}_{ijt}}}}$$

(5)

In Equation (5), $h{s}_{jt}$ denotes horizontal spillovers. ${\rho }_{ijt}$ is the foreign capital share of firm i in industry j in year t, which is expressed by the share of foreign ownership of firm i. $y_{ijt}$ denotes the main business revenue of firm i, and ${\theta }_j$ is the set consisting of all firms belonging to industry j in the sample. The horizontal spillover effect increases along with the increase in the foreign capital market share in the industry and the proportion of foreign shareholding in each enterprise.

2.

Vertical spillovers

Vertical spillovers are usually categorized into forward spillovers (fs) and backward spillovers (bs). Forward spillovers occur when foreign suppliers provide inputs to downstream domestic customers. Forward spillovers are calculated as follows:

$$f{s}_{jt}={\displaystyle \sum }_{j\ne k}{\alpha }_{jkt}\times h{s}_{jt}$$

(6)

where ${\alpha }_{jkt}$ is the forward correlation coefficient, derived from the values in each column of the direct consumption coefficients of the 2012 Input–Output Table of China. It is the proportion of the inputs purchased by industry j from the upstream industry k in year t to the total inputs of industry j.

Backward spillovers occur when local suppliers provide inputs to downstream foreign customers. Backward spillovers are calculated as follows:

$$b{s}_{jt}={\displaystyle \sum }_{j\ne k}{\sigma }_{jkt}\times h{s}_{jt}$$

(7)

where ${\sigma }_{jkt}$ is the backward correlation coefficient, derived from the values in each row of the direct consumption coefficients. It represents the proportion of outputs sold by industry j to downstream industry k in year t to the total outputs of industry j.

3.2.3. Threshold Variable

Absorptive capacity of the enterprise (absorb): Absorptive capacity is a multidimensional concept primarily discussed through three-dimensional and four-dimensional perspectives. Cohen and Levinthal [47] defined absorptive capacity as a firm’s ability to identify, assimilate, and commercially apply new external knowledge. Zahra and George [48] defined absorptive capacity as the dynamic process by which an organization acquires, assimilates, transforms, and exploits external advanced knowledge. Due to the practical challenges in clearly distinguishing between the transformation and exploitation of new knowledge in business practices [49], this study adopts the three-dimensional framework of absorptive capacity, categorizing it into knowledge acquisition, assimilation, and transformation and integration capacity. Zhen and Tang [50] categorized absorptive capacity into three dimensions and selected corresponding variables to measure these dimensions. Drawing on the practice of Zhen and Tang [50], this study constructs an index system from the aspects of acquisition, assimilation, and transformation and integration capacity (See

Table 1. A comprehensive evaluation indicator system for the absorptive capacity of the enterprise.

1st-Level Indicators	2nd-Level Indicators	3rd-Level Indicators	Variable Definition
Absorptive capacity of the enterprise	Acquisition capacity	Degree of diversification	Herfindahl index
		Degree of services concentration in the province where the enterprise is located	Employment in the service sector in the province where the enterprise is located/national employment in the service sector
	Assimilation capacity	Level of human capital	Average employee salary
		Internal management efficiency	The ratio of administrative expenses to operating revenue
	Transformation and integration capacity	Technological gap	TFP of the enterprise/maximum TFP of enterprises in the sector
		R&D level	The proportion of intangible assets in the total assets of the enterprise

). These three aspects are distinct yet interconnected, and the synergistic effect between them is the key to the enhanced absorptive capacity. Then, the factor analysis method is employed to aggregate the quantifiable indicators of the three aspects, extract common factors, and calculate the comprehensive score of absorptive capacity for each enterprise.

3.2.4. Control Variables

This study selects enterprise age (age), enterprise capital intensity (kl), return on assets (roa), enterprise profitability (profit), asset-to-liability ratio (debt) and the degree of market competition (hhi) as control variables. The specific variable definitions and descriptions are presented in

Table 2. Variable definitions and descriptions.

Variable Type	Variable Name	Variable Symbol	Variable Definition
Explained variable	High-quality development of enterprises	TFP	The logarithm of total factor productivity calculated by the LP method
Explanatory Variables	Horizontal spillovers	hs	The proportion of the main business revenue of MNC to the total main business revenue of the industry
	Forward spillovers	fs	The proportion of the input purchased from upstream foreign firms to the total input of the sector
	Backward spillovers	bs	The proportion of the output sold to downstream foreign firms to the total output of the sector
Threshold Variable	Absorptive capacity of the enterprise	absorb	See Table 1
Control variables	Enterprise age	age	ln(year of observation-year of establishment + 1)
	Enterprise capital intensity	kl	net book value of fixed assets/the number of employees
	Return on assets	roa	Net profit/total assets
	Enterprise profitability	profit	Net profit/total operating revenue
	Asset-to-liability ratio	debt	Total liabilities/total assets
	Degree of market competition	hhi	Herfindahl index of the two-digit industry

.

3.3. Data Source and Descriptive Statistics

This study selected A-share listed service enterprises in China from 2012 to 2019 as samples, excluding samples of the financial sector, special treatment (ST and *ST) listed companies, and samples with missing values. Finally, 535 listed service enterprises with 4280 observations were obtained. Enterprise data were collected from CSMAR and Wind databases. Industry data were collected from the China Statistical Yearbook of the Tertiary Industry and the statistical yearbooks of each province. The correlation coefficients were obtained from the Input–Output Table of China from the National Bureau of Statistics of China. All continuous variables were winsorized at 1% and 99%. The descriptive statistics of the variables are shown in

Table 3. Descriptive statistics.

Variable	N	Mean	SD	Min	Max
TFP	4280	9.3284	1.2793	4.4579	13.5152
hs	4280	0.0480	0.0455	0.0002	0.2021
fs	4280	0.0280	0.0167	0.0050	0.0749
bs	4280	0.0167	0.0140	3.88 × 10−6	0.0706
age	4280	2.9742	0.2919	1.0986	3.9890
absorb	4280	1.01× 10−10	0.5816	−1.6190	7.8580
kl	4280	0.0090	0.0825	0.00003	3.0448
roa	4280	0.0383	0.0721	−0.5637	1.8525
profit	4280	0.1310	2.2924	−15.7422	109.7486
debt	4280	0.4613	0.2491	0.0103	8.2564
hhi	4280	0.1716	0.1981	0	1

.

4. Results and Discussion

4.1. Benchmark Regression Results

The variables were first tested for multicollinearity, and the results show that the variance inflation factor (VIF) is 1.39, indicating no serious multicollinearity concerns in the regression model. This study adopts the double fixed effect model with both individual and time-fixed effects, and the regression results are shown in

Table 4. Benchmark regression results.

Variables	(1) TFP	(2) TFP
hs	0.8668 *** (2.83)	0.9026 *** (2.94)
bs	5.4052 *** (2.93)	5.3563 *** (2.98)
fs	−4.8030 *** (−3.33)	−4.9436 *** (−3.51)
age		0.7056 *** (3.00)
roa		1.3529 *** (4.55)
kl		0.4926 *** (5.30)
profit		−0.0322 *** (−4.09)
hhi		−0.0523 (−0.38)
debt		0.4617 (1.28)
Constant	9.3309 *** (256.98)	6.9792 *** (10.96)
Firm FE	YES	YES
Year FE	YES	YES
Observations	4280	4280
R-squared	0.8652	0.8709

Notes: *** p < 0.01; t-values in parentheses adjusted for robust standard errors, with the same below.

. The empirical results without any control variables can be observed in column (1). Column (2) measures the empirical results of adding control variables and controlling for individual and time effects. As shown in column (2), horizontal and backward spillovers have a significant positive impact. This indicates that at the intra-industry level, FDI enhances the high-quality development of domestic enterprises through the demonstration, competition, and personnel mobility effect. At the inter-industry level, upstream domestic enterprises can improve their TFP through linkages with downstream MNC clients. The coefficient of backward spillovers is higher and significant at the 1% level, suggesting a more pronounced backward spillover effect. Hypothesis 1a and Hypothesis 1b are verified.

While most existing studies on Chinese manufacturing firms have concluded that FDI has positive forward spillovers [9,10], this study finds that the coefficient of forward spillovers is significantly negative. Evidence for insignificant or negative forward spillovers in services can be found in firm-level panel data studies for some transition or developing economies [51,52]. The possible explanations for negative forward spillovers are as follows. First, Miozzo and Grimshaw [53] argued that FDI forward linkages in less developed countries are influenced by institutional and economic characteristics of the host economy. For a long time, the opening-up pace of China’s service industry has lagged behind that of the manufacturing industry. Market access and foreign equity restrictions in upstream service sectors have constrained FDI inflow, limiting downstream enterprises’ access to high-quality service intermediates, thereby generating negative forward spillover effects. Second, the high-quality inputs provided by upstream MNCs may also entail higher prices, and domestic firms may suffer from the negative impact of high costs if they fail to derive sufficient benefits from them [54]. Third, domestic firms also need to improve their service quality and efficiency to adapt to the high-quality inputs provided by upstream MNCs. If domestic firms have limited absorptive capacity and lack appropriate technical and managerial staff, it may be difficult to incorporate high-quality inputs from MNCs into their production processes [34].

4.2. Robustness Test

To verify the robustness of the above findings, this study conducts a series of robustness tests by replacing the explained variable, changing the sample size, and controlling over endogeneity.

4.2.1. Replace the Explained Variable

This study re-estimates the model by changing the measurement method of the explained variable. The total factor productivity of sample enterprises is recalculated using the ACF method, and the results are presented in column (1) of

Table 5. Robustness test.

Variables	Replace the Explained Variable	Change the Sample Size	IV1	IV2
	(1) TFP(ACF)	(2) TFP	(3) TFP	(4) TFP
hs	0.7913 *** (2.77)	0.7558 ** (1.99)	2.0416 ** (2.12)	1.3300 *** (3.58)
bs	4.4478 *** (2.59)	5.8330 *** (2.60)	8.6059 *** (4.33)	6.3762 *** (3.51)
fs	−3.6115 *** (−2.67)	−4.8928 *** (−2.81)	−3.7550 *** (−3.95)	−3.0570 *** (−3.18)
age	0.5715 *** (2.78)	0.8038 *** (2.95)	1.9414 *** (11.42)	2.0322 *** (24.64)
roa	1.1129 *** (4.46)	1.2821 *** (3.74)	1.8292 *** (7.41)	1.3890 *** (4.65)
kl	0.9923 *** (4.82)	0.4308 *** (4.34)	0.4547 *** (5.16)	0.5489 *** (5.37)
profit	−0.0286 *** (−4.00)	−0.0300 *** (−3.92)	−0.0281 *** (−4.44)	−0.0332 *** (−4.16)
hhi	0.1763 (1.32)	−0.1762 (−1.03)	−0.1828 (−1.35)	−0.0530 (−0.38)
debt	0.4728 * (1.70)	0.3592 (1.04)	1.3024 *** (8.08)	0.4307 (1.26)
Constant	5.8186 *** (10.16)	6.6727 *** (8.73)
Firm FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Observations	4280	3056	3745	4280
LM statistic			514.987 [0.00]	193.551 [0.00]
Wald F statistic			270.298 [16.38]	881.962 [16.38]
R-squared	0.8669	0.8564	0.2828	0.2701

Notes: * p < 0.1, ** p < 0.05, and *** p < 0.01; the LM statistic and Wald F statistic are used for testing the under-identification and weak-identification of instrumental variables, respectively.

. The regression results are consistent with those of the benchmark regression.

4.2.2. Change the Sample Size

China’s municipalities have relatively independent decision-making and management authority, so the development of FDI may be different from other provinces. Therefore, this study re-estimates the model by eliminating the sample of municipalities. As can be seen from column (2) of Table 5, the sign and significance level of the regression coefficients of the three FDI spillover variables do not change fundamentally compared with the benchmark regression, which again proves the robustness of the regression results.

4.2.3. Endogenous Analysis

Instrumental variable analysis and a two-step GMM approach are employed to address endogeneity in this study. First, the one-phase lagged explanatory variable is used as an instrumental variable, and the regression results are shown in column (3) in Table 5. Second, following the approach of Zhang et al. [55], this study uses the interaction term between the reciprocal of the distance from the enterprise’s city to the coastline and the foreign capital entry intensity in various service sectors as an instrumental variable. The regression results are shown in column (4) in Table 5. The results of both the Kleibergen-Paap rk LM test and Wald F test reject the null hypothesis, which validates the reliability of the IV estimation. After considering potential endogeneity problems using instrumental variables, the empirical research conclusions of this study are consistent with the findings acquired above and remain robust.

4.3. Heterogeneity Analysis

To further discuss the heterogeneity of the impact of FDI spillovers, the benchmark model is regressed in groups according to the heterogeneous characteristics of firms, such as geographical locations, intensity of enterprise factor inputs, size of the enterprise, ownership of the enterprise, and industry classification.

4.3.1. Analysis Based on Different Geographical Locations

To further explore whether there are regional differences, the sample enterprises are categorized into the eastern, central, and western regions based on the geographical location of the enterprises. The regression results presented in column (1) and column (2) of

Table 6. Heterogeneity test I.

Variables	Eastern Region	Central and Western Regions	Capital-Intensive	Labor-Intensive
	(1) TFP	(2) TFP	(3) TFP	(4) TFP
hs	1.0645 *** (3.25)	−0.1972 (−0.28)	0.9670 ** (2.23)	0.0949 (0.24)
bs	3.5982 *** (2.01)	14.1987 ** (2.38)	3.7037 * (1.72)	2.5709 (1.06)
fs	−4.9552 *** (−3.21)	−5.8049 * (−1.72)	−4.0623 ** (−2.44)	−2.4709 (−1.12)
Control variables	YES	YES	YES	YES
Constant	6.0470 *** (7.65)	8.2521 *** (8.28)	8.4925 *** (10.44)	6.9246 *** (7.23)
Firm FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Observations	3264	1016	2140	2140
R-squared	0.8791	0.8611	0.9140	0.8996

Notes: * p < 0.1, ** p < 0.05, and *** p < 0.01.

indicate that the coefficient of FDI horizontal spillovers in the eastern region is significantly positive but not significant in the central and western regions. The possible reason is that regions with advanced service development demonstrate a stronger capacity to attract foreign investment, which in turn contributes to the formation of service industry clusters characterized by the coexistence of domestic and foreign capital [56]. The existence of service industry clusters reduces the likelihood of temporal lag and spatial attenuation in the diffusion of knowledge and technology, thereby enabling domestic service enterprises to assimilate technological and managerial practices from service MNCs more effectively. Therefore, positive FDI horizontal spillovers are more likely to occur in the eastern region, where the development level of the service industry is relatively high.

4.3.2. Analysis Based on Different Intensity of Factor Inputs

This study classifies enterprises with higher capital intensity than the median as capital-intensive enterprises and those with a lower intensity as labor-intensive enterprises. The results in columns (3) and (4) of Table 6 show that, compared to labor-intensive enterprises, the impact of FDI on capital-intensive enterprises is more pronounced, not only in terms of the facilitating effect of horizontal and backward spillovers but also in terms of the inhibiting effect of forward spillovers. The research of Chen and Chen [57] indicated that FDI has a positive effect on employment in capital-intensive producer services such as transportation, leasing, and business services. This could be explained by the fact that for domestic capital-intensive enterprises that require high levels of technological expertise and capital investment, foreign investment not only alleviates the financing constraints but also introduces technological and managerial models that serve as valuable references, enabling domestic capital-intensive enterprises to achieve efficient capital operation and technological innovation.

4.3.3. Analysis Based on Different Size

The size of the enterprise is measured by taking the natural logarithm of its total assets. Enterprises with total assets above the median are classified as large-scale enterprises, while those below the median are classified as small and medium-sized enterprises (SMEs). The results in columns (1) and (2) of

Table 7. Heterogeneity test II.

Variables	Large-Scale	SMEs	SOEs	Non-SOEs	Producer Services	Consumer Services
	(1) TFP	(2) TFP	(3) TFP	(4) TFP	(5) TFP	(6) TFP
hs	−0.1501 (−0.41)	1.0320 *** (2.58)	−0.2919 (−0.67)	0.0149 (0.04)	0.9039 *** (2.76)	1.0893 (0.33)
bs	2.3768 (1.27)	4.7424 **(2.08)	3.9298 (1.57)	4.4539 * (1.77)	6.0792 *** (3.20)	−3.3653 (−0.32)
fs	−1.6855 (−1.12)	−5.7320 *** (−2.83)	−3.1428 * (−1.73)	−3.8143 (−1.57)	−5.3794 *** (−3.41)	−8.6322 (−0.68)
Control variables	YES	YES	YES	YES	YES	YES
Constant	10.2158 *** (14.90)	4.7316 *** (5.93)	8.8284 *** (14.37)	6.9907 *** (6.22)	7.1798 *** (10.45)	7.0510 *** (5.13)
Firm FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
Observations	2140	2140	2040	2240	3808	472
R-squared	0.9066	0.8374	0.8766	0. 8653	0.8748	0.7926

Notes: * p < 0.1, ** p < 0.05, and *** p < 0.01.

show that compared to large-scale enterprises, horizontal spillovers and backward spillovers have a more significant positive effect on the high-quality development of SMEs. It is generally held that large-scale enterprises tend to increase R&D investment when facing competitive pressure from foreign counterparts, thereby enhancing corporate innovation efficiency. Nevertheless, the regression results indicate that the enterprise size of service enterprises inhibits the positive impact of FDI spillovers. This could be explained by the fact that large-scale enterprises are more likely to adopt conservative innovation strategies due to the presence of “organizational inertia” and “organizational rigidity”, thereby persisting in existing technological trajectories to achieve directed technological innovation, whereas SMEs are characterized by high flexibility and specialization, and are better positioned in disruptive technological innovation [58]. Therefore, the impact of spillovers on the high-quality development of SMEs is more significant.

4.3.4. Analysis Based on Different Ownership

Previous studies have found that ownership disparities influence the corporate capacity to procure resources and formulate economic decisions [59]. Chinese enterprises of different ownership types exhibit significant disparities in government policy support, financing capacity, and operational performance, leading to differentiated efficiency in foreign capital utilization. The sample enterprises are categorized into state-owned enterprises (SOEs) and non-state-owned enterprises (non-SOEs) based on ownership. The regression results for SOEs and non-SOEs are reported in columns (3) and (4) of Table 7, respectively. The regression results indicate that for non-SOEs, the impact of backward spillovers is more pronounced. Although the “policy preferences” granted to SOEs can provide resource advantages, these institutional benefits and privileged allocations may paradoxically diminish their motivation to establish business partnerships with MNCs [60]. Chen et al. [61] measured the share of private enterprises in the supply chains of MNCs and concluded that MNCs prefer to choose private enterprises as their suppliers because private enterprises have stronger learning ability, better absorption capacity, higher decision-making efficiency, and lower transaction costs.

4.3.5. Analysis Based on Different Industry Classification

The spillover effects of FDI may also vary across different sub-sectors within the service industry. Existing studies on China’s service industry typically categorize it into two primary types: producer services and consumer services. Consequently, the sample enterprises are classified into producer service enterprises and consumer service enterprises. The results in columns (5) and (6) of Table 7 show that compared to consumer services, the growing presence of FDI in producer services can significantly promote horizontal and backward spillover effects. This may be attributed to the fact that greater openness in producer services directly provides enterprises with more accessible and specialized services such as finance, warehousing and logistics, IT infrastructure, industrial leasing, and management and training support, thereby facilitating high-quality development and amplifying FDI spillovers.

4.4. Test of Threshold Effect

Using the threshold panel regression model proposed by Hansen [44], the study estimates the threshold values of absorptive capacity for each spillover channel. As can be seen from

Table 8. Test of threshold effect.

Key Explanatory Variables	Model	F-Value	p-Value	Critical Value
				1%	5%	10%
hs	Single threshold	425.53	0.0000	34.6619	25.4778	21.4294
	Double threshold	97.13	0.0000	26.7032	21.3751	18.7991
	Triple threshold	60.95	0.5533	136.0193	113.9588	103.6602
bs	Single threshold	499.16	0.0000	26.2645	17.8501	15.2914
	Double threshold	156.09	0.0000	23.5636	16.5947	14.5885
	Triple threshold	129.10	0.5100	213.0167	183.5319	168.3923
fs	Single threshold	578.28	0.0000	20.1771	15.3502	13.4104
	Double threshold	189.62	0.0000	23.7336	16.1647	13.6488
	Triple threshold	115.45	0.5233	207.9393	176.2673	159.5902

, when the threshold effect test is conducted with horizontal spillover as the core variable, the F values of the double-threshold and triple-threshold tests are 97.13 and 60.95, and the p values are 0.0000 and 0.5533, indicating the existence of the double-threshold effect. Similarly, when the threshold effect test is conducted with forward and backward spillover as the core variables, the single threshold and double threshold also pass the significance test, respectively. Therefore, this study adopts the double-threshold model to investigate the nonlinear impact of FDI spillovers on the high-quality development of service enterprises. The specific threshold estimates of absorptive capacity are shown in

Table 9. Test of threshold estimates for absorptive capacity.

Key Explanatory Variables	Model	Threshold Estimate	95% Confidence Interval
hs	Single threshold	−0.6449	[−0.6666, −0.6135]
	Double threshold	−0.3659	[−0.3815, −0.3469]
		−0.2976	[−0.3198, −0.2512]
bs	Single threshold	−0.4371	[−0.4979, −0.4252]
	Double threshold	−0.4371	[−0.4542, −0.4252]
		−0.0742	[−0.0846, −0.0639]
fs	Single threshold	−0.5127	[−0.5342, −0.4979]
	Double threshold	−0.4371	[−0.4542, −0.4252]
		−0.0639	[−0.0914, −0.0535]

. The double-threshold effect of absorptive capacity implies that under different levels of absorptive capacity, the impact of FDI spillovers demonstrates different directions and intensities. Hypothesis 2 is verified.

Column (1) in

Table 10. The test results of the double threshold effect of absorptive capacity.

Variables	(1) TFP	(2) TFP	(3) TFP
hs		1.1188 *** (3.10)	1.0024 *** (2.81)
bs	4.5777 *** (3.81)		7.0427 *** (4.28)
fs	−2.3747 *** (−3.75)	−2.8002 *** (−3.44)
age	1.8830 *** (15.48)	1.8159 *** (15.48)	1.8074 *** (15.92)
roa	1.1775 *** (4.67)	1.1667 *** (4.81)	1.0889 *** (4.63)
kl	0.5829 *** (11.91)	0.5017 *** (9.12)	0.5302 *** (11.38)
profit	−0.0275 *** (−3.15)	−0.0293 *** (−4.92)	−0.0271 *** (−5.66)
hhi	−0.0252 (−0.12)	−0.1367 (−0.66)	−0.0399 (−0.20)
debt	0.4126 (1.18)	0.3964 (1.18)	0.3882 (1.22)
constant	3.4222 *** (9.02)	3.6652 *** (10.05)	3.6991 *** (10.43)
$\mathit{absorb}\le {\theta }_1$	−6.1414 *** (−6.64)	−17.6126 *** (−6.41)	−24.0107 *** (−10.83)
${\theta }_1$$<\mathit{absorb}\le {\theta }_2$	−1.3181 ** (−2.48)	2.6094 (1.36)	−6.8863 *** (−6.22)
$\mathit{absorb}>{\theta }_2$	4.0548 *** (9.02)	15.2902 *** (7.02)	2.9882 *** (3.14)
R-squared	0.1559	0.1960	0.2155

Notes: ** p < 0.05, and *** p < 0.01.

provides the estimated result with horizontal spillovers as the core explanatory variable and absorptive capacity as the threshold variable. When the absorptive capacity is in the low range (absorb ≤ −0.3659) and the middle range (−0.3659 < absorb ≤ −0.2976), the coefficients of FDI horizontal spillovers are −6.1414 and −1.3181, respectively, and they have passed the significance test at the 1% level. When the absorptive capacity is in the high range (absorb > −0.2976), the coefficient of FDI horizontal spillovers is 4.0548 and passes the 1% significance test. It can be seen that as absorptive capacity improves, the impact of horizontal spillovers on the high-quality development of service enterprises gradually increases. Column (2) and column (3) in Table 10 present the estimated result with backward and forward spillovers as core explanatory variables, respectively. Similar to the horizontal spillovers, the positive backward and forward effects are detected if the absorptive capacity is in the high range. A higher absorptive capacity of an enterprise can promote its active acquisition and assimilation of FDI spillovers, achieving the transformation and integration of spillovers. It should be noted that among the three spillovers, a positive influence of forward spillovers demands that enterprises have a higher absorptive capacity. This also explains why the coefficient of forward spillovers in the benchmark regression is significantly negative. This verifies that the absorptive capacity of domestic service enterprises is not sufficient to utilize the high-quality inputs from upstream MNCs effectively.

5. Research Conclusions and Policy Implications

5.1. Conclusions

This study selects A-share listed service enterprises in China from 2012 to 2019 as the research sample to investigate the effects and mechanism of FDI spillovers on the high-quality development of service enterprises. The main research conclusions are as follows: (1) Horizontal and backward spillovers have a significant positive impact on the high-quality development of service enterprises, while forward spillovers have a significant negative impact. (2) Heterogeneity analysis demonstrates that FDI horizontal spillovers have a more pronounced impact on service enterprises in the eastern region. Compared to labor-intensive, large-scale, and consumer service enterprises, horizontal spillovers and backward spillovers have a more significant positive effect on capital-intensive producer service enterprises and SMEs. Compared to SOEs, the impact of backward spillovers is more pronounced for non-SOEs. (3) The further threshold regression model finds a double-threshold effect of absorptive capacity in the impact of FDI spillovers on the high-quality development of service enterprises. The direction and intensity of the three spillovers depend on the absorptive capacity of domestic enterprises. Furthermore, compared with the horizontal spillover effect, the vertical spillover effect, especially the forward spillover effect, requires that service enterprises possess a higher absorptive capacity.

5.2. Policy Implications

Based on the research findings, this study proposes the following policy recommendations. Firstly, the government should make greater efforts to attract and utilize foreign investment and steadily and continuously expand market access in the service industry. The government should also intensify efforts to attract investment in modern service sectors and continuously enhance the level of high-standard opening-up of the service industry. Policymakers should forge a fairer and more competitive market operating environment to guide and create incentives for effective linkages between MNCs and domestic enterprises, facilitating the establishment of vertical chains and the entry of domestic enterprises into the global value chain. Given the negative impact of forward spillovers, the government should increase efforts to attract foreign investment in upstream producer services, enabling more foreign-invested service providers to enter sectors such as finance, securities, R&D, intermediary consulting, marketing, and information networks. This would allow downstream enterprises to access a greater variety of intermediate service products at lower costs, thereby enhancing the forward spillover effects of FDI. In recent years, China has continuously revised the “Special Administrative Measures for Foreign Investment Access (Negative List)”, gradually eliminating or relaxing foreign equity restrictions in producer services. Meanwhile, more open foreign investment policies have been piloted in free trade zones, permitting foreign investors to launch innovative businesses in fields like data services and scientific R&D.

Secondly, service enterprises as suppliers should strive to establish closer relations with MNCs and provide differentiated and diverse intermediate service items. Meanwhile, enterprises need to integrate the objective of conforming to the supplier standards of MNCs with their long-term development goals and formulate the overall strategic development plan of the enterprises. Through product innovation, market innovation, technological innovation, resource allocation innovation, and institutional innovation, they can realize their high-quality development and enhance the capacity for sustainable growth and continuous value creation.

Thirdly, service enterprises should enhance their absorptive capacity to better utilize the spillover effects of FDI and align their absorptive capacity with the dynamic changes in the global value chain. Enterprises should attach significance not only to the identification and acquisition of external knowledge but also to the transformation and utilization of such knowledge. Firstly, it is necessary to enhance the capacity to acquire external knowledge that holds potential value for the competitive edge of enterprises and expand the diversity of enterprise knowledge. Secondly, it is essential for service enterprises to elevate human capital and management efficiency. Service enterprises need to intensify their investment in human capital and cultivate and introduce innovative talents to increase the stock level of human capital in the enterprises. Meanwhile, they should resort to digital technologies to enhance the management efficiency of the enterprises. Thirdly, service enterprises should increase their R&D investment, enhance their independent innovation capabilities, and achieve breakthroughs in service content, form, and organizational structure. A successful case is China’s JD Logistics. By introducing the intelligent warehouse management and global supply chain management expertise of foreign-funded enterprises such as Walmart and DHL and adapting these practices to the unique characteristics of China’s e-commerce market through localization, JD Logistics has developed the “211 time-delivery” service (guaranteeing delivery by 11 AM for orders placed the previous day, or by 11 PM for same-day orders) tailored to meet Chinese consumers’ demands. This initiative has significantly enhanced logistics efficiency, and the proportion of cross-border business revenue has continued to grow. Additionally, enterprises can rationally leverage the spillover effects of FDI based on their varying absorptive capacities. Domestic service enterprises with a relatively strong absorptive capacity can employ the high-quality intermediate service products offered by upstream MNCs, thereby elevating their innovation capabilities. Service enterprises with a relatively weak absorptive capacity can achieve high-quality development by imitating and learning from their foreign counterparts, actively attracting the labor force with working experience in foreign enterprises, and carrying out extensive technical collaboration and knowledge and information sharing with downstream foreign enterprises.

5.3. Limitations and Future Research

This study has some limitations, which could be addressed in future research. First, China only collects micro-level enterprise data during economic censuses in individual years, making it difficult to obtain continuous and comprehensive data on service enterprises other than listed enterprises. Therefore, this study focuses on listed companies, which may limit the generalizability of the research findings. In the future, if more data become accessible, the dataset could be expanded beyond listed companies to encompass a broader range of service enterprises, particularly small and medium-sized enterprises. Second, this study is restricted to a sample of service enterprises. Future research could extend to comparative studies with manufacturing enterprises to explore industry-specific spillover effects, aiding targeted FDI policy design. Finally, given the variations in policy environments, economic and social development, and technological capabilities across different countries, future research could include economies with expanding service sectors and strong FDI inflows to enhance the generalizability of the research findings.