Can the Opening of High-Speed Railway Restrain Corporate Financialization?

Abstract

Under the background of the economy “shifting from real to virtual”, how to guide real enterprises to return to their main businesses has become an urgent problem to be solved in the stage of microeconomic bodies moving toward high-quality development. Based on the perspective of the opening of high-speed railway (HSR), this paper builds a time-varying DID model to systematically test the relationship between HSR and corporate financialization by matching the data of HSR among 286 cities in China with the data of A-share non-financial companies listed in the Shanghai and Shenzhen stock markets. The results showed that there is a significant negative correlation between HSR and corporate financialization, i.e., the opening of HSR can significantly inhibit corporate financialization. This conclusion still holds after a series of robustness tests. Mechanism analysis found that the opening of HSR can restrain corporate financialization through “crowding-out effects” and “liquidity replenishment effects”. Heterogeneity analysis showed that the inhibitory effect of HSR on corporate financialization is more significant in large enterprises, fiercely competitive industries, and cities with high initial transportation resources. This paper deeply explored the relationship between the opening of HSR and corporate financialization, which not only enriches the existing literature but also provides a useful reference for guiding real enterprises to actively return to their main businesses.

1. Introduction

Under the new normal, the continuing international trade disputes together with the domestic economic downturn narrow the profit margin of real enterprises in China. Meanwhile, the financial industry has made the return on financial assets show an obvious countercyclical upward trend. On this condition, the profit-seeking nature of capital drives real enterprises to deviate from their main business and to invest a large number of funds into the financial system, thus forming the “hollowing out” of the manufacturing industry, that is, the financialization of real enterprises (Du et al., 2020) [1]. The lack of suitability between the manufacture and the finance will lead to a structural imbalance between the real economy and the virtual economy, which becomes a stumbling block impeding the sustainable development of both the manufacturing and financial industries. Furthermore, corporate financialization will blur the boundary between the real and virtual sectors, intensifying the risk linkage between the two, thus increasing the probability of outbreaks of systemic financial risks (Bhaduri, 2011) [2]. In the face of severe reality, President Xi Jin-Ping during the 20th CPC National Congress pointed out “persisting in the principle of putting the focus of economic development on the real economy” and “strengthening modern financial supervision to prevent the occurrence of systematic financial risks”. Under this background, it has become an urgent problem to be solved, for both the government and academia to deeply explore the motivation of real entities’ financialization so as to actively guide them to return to their main business in the stage of microeconomic entities moving toward high-quality development.

In essence, corporate financialization is an investment decision that decides how to build the investment portfolio with physical assets and financial assets. The literature on the impact of transportation infrastructure on enterprise investment decisions has achieved fruitful results. Especially in the context where China’s high-speed railway (HSR) construction gained rapid growth, the impact of “space-time compression effects” caused by HSR on enterprise investment decisions has become the focus of scholars. Specifically, the existing literature discusses the impact of HSR on cross-regional investment (Ma et al., 2020) [3], venture capital investment (Long et al., 2017) [4], geographical expansion (Wang et al., 2020) [5], and location choice of investment (Zhuang et al., 2021) [6], etc. However, it should be noted that there is no research directly focusing on HSR and corporate financialization. Does the opening of the HSR affect corporate financialization? What is the theoretical mechanism between the two? The answers to the above questions are not only conducive to preventing the real entities from being financialized, but also provide a useful reference for policymakers to scientifically plan the construction of HSR and further activate the catalytic role of HSR in the process of moving toward high-quality development.

Therefore, this paper builds a multi-period DID model to empirically test the relationship between the opening of HSR and corporate financialization by matching the data of HSR on the city level with the data of nonfinancial A-share companies listed in Shanghai and Shenzhen stock market from 2007 to 2020. The results showed that the opening of HSR can significantly inhibit corporate financialization. This conclusion still holds after a series of robustness tests. Mechanism analysis found that the opening of HSR can inhabit corporate financialization through “crowding-out effects” and “liquidity replenishment effects”. Furthermore, the inhibitory effect of the opening of HSR on corporate financialization is more significant in large-scale enterprises, fierce competitive industries, and cities with high initial transportation resources.

The possible marginal contributions of this paper are as follows: Firstly, since there is no research directly focusing on the opening of HSR and corporate financialization, we innovatively bring the two into the same analytical framework and empirically explore the relationship between the two, which fills the gap in this research branch. Secondly, we also interpret HSR’s heterogeneous impact on corporate financialization from different dimensions, including firm-level, industry-level, and city-level, which describes the relationship between the two in a panoramic way. Thirdly, different from prior literature that analyzed corporate financialization either from “reservoir” theory or investment substitution theory, we construct a comprehensive theoretical framework combining both theories to uncover the possible mechanism between HSR and corporate financialization, which not only enriches the study on the economic consequences of HSR opening, but also expands the research scope for the study on the motivation of corporate financialization.

This paper focuses on China’s context for two reasons: on the one hand, as the largest transitional economy around the world, dealing with “shifting from real to virtual” is not only helpful for guiding real enterprises to return to their main business but also meaningful for academia deeply learning the essence of financialization in a transitional economy. On the other hand, the rapid growth of HSR in China has provided an excellent opportunity for researchers to empirically test the economic consequences of HSR.

This paper is structed as follows: Section 2 presents the literature review and mechanism analysis. Section 3 describes the data and methodology. Section 4 documents the empirical results. Section 5 concludes and highlights discussions.

2. Literature Review and Mechanism Analysis

2.1. Literature Review

2.1.1. Research on Corporate Financialization

Financialization refers to the increasing contributions of the financial sector to the national economy (Palley, 2008) [7], which reflects the process by which the mode of profit accumulation changes from production and exchange of surplus value to capital appreciation (Orhangazi, 2007) [8].

Previous studies have explored the motivations of corporate financialization in an extensive way. At the macro level, there are many studies that have had a profound influence (Krippner 2005; Orchangazi, 2008; Demir, 2009; Zhang and Zhang, 2016; Peng et al., 2018; Hu et al., 2017) [8,9,10,11,12,13]. Krippner (2005) [13] believed that the decline in the profit margin of traditional manufacturing industries is the main reason why real entities are willing to engage in financial investments (Zhang and Zhang) [11]. Specifically, the large profit gap between financial assets and physical assets drives manufacturing companies to invest more capital into financial sectors so that physical investments have been crowded out (Orhangazi, 2007) [8]. Such firms attach more importance to capital operations and ignore production activities, resulting in a continuing rise in the proportion of financial profits (Krippner, 2005) [13], which is a typical feature of being financialized. After analyzing the portfolios of nonfinancial enterprises in Argentina and other countries, Demir (2009) [12] found that in addition to yield differences, dealing with macroeconomic uncertainties is one of the most important reasons for enterprises to hold financial assets. Peng et al. (2018) [10] found the same result; they argued that enterprises would reduce their holdings of speculative financial assets in consideration of expected risks when facing severe economic uncertainty. Since then, the research at the macro level is more about the impact of external policy shocks on corporate financialization, such as monetary policy (Yang et al., 2019; An et al., 2022) [14,15], macro-prudential policy (Ma and Chen, 2020) [16], and industrial policy (Bu et al., 2020) [17], etc.

The research at the microlevel mainly focuses on the following aspects. First, there is the influence of management characteristics on corporate financialization. Studies in this branch are mostly based on the upper echelon theory and the branding theory to explore how the personal experience of the managers affects corporate financialization (Du et al., 2019; Du and Zhou, 2019) [1,18]. Specifically, the special experience of senior executives in the “sensitive period” will leave a “mark” on them. This “mark” will continue to affect their cognition over time, and then affect the investment decisions of enterprises. Relevant research has discussed the relationship between corporate financialization and CEO financial background (Du et al., 2019) [1], executive academic background (Du and Zhou, 2019) [18], and international directors (Dai and Wang, 2021) [19], etc. Second come corporate social responsibility and financialization. Altruism and egoism are two opposing concepts of social responsibility that also have different impacts on corporate financialization. Egoists regard fulfilling social responsibilities to obtain scarce external resources. For example, Gu et al. (2020) [20] found that enterprises with well-performed social responsibilities can alleviate their financing constraints and thus increase their degree of financialization. The altruists pay more attention to the demands of external stakeholders and the sustainable development of enterprises, and this concept has developed into a management tool to inhibit corporate financialization (Liu et al., 2019) [21]. Third, we have corporate governance and financialization. Research in this branch is based on the principal-agent theory to investigate the impact of improved corporate governance on corporate financialization. For example, Wang and Huang (2020) [22] found that high-quality internal control can strengthen the supervision of the management and inhibit the management’s over-allocation of financial assets due to self-interest motivation. Jiang et al. (2022) [23] found that when private shareholders, improve their shareholdings and gain more power, they can play a monitoring role on controlling shareholders, which can significantly inhibit corporate financialization.

2.1.2. Research on Economic Consequences of Opening HSR

Research on examining the economic consequences of HSR has emerged in large numbers in recent years. Specifically, studies are mainly carried out from the following aspects: first, HSR and urban innovation. Most of the research in this direction examines the promotional effect of HSR on urban innovation from the perspective of both factor mobility and knowledge spillover (Yang et al., 2021; Ji and Yang, 2020; Chen et al., 2019; Ye et al., 2020) [24,25,26,27]. From the view of factor mobility, HSR breaks the market segmentation and promotes the cross-regional flow of innovative factors so that the high wage returns in central cities will attract the influx of the senior labor force, thus prompting the innovation of local enterprises (Ji and Yang, 2020; Chen et al., 2019) [25,26]. From the perspective of knowledge spillover, the economic agglomeration effect generated by the opening of HSR has promoted the mutual integration of knowledge among various industries, which is conducive to the generation of new knowledge and creating knowledge spillover (Ye et al., 2020) [24]. The second aspect is HSR and the cross-regional capital flow. The research in this branch discusses whether the opening of HSR promotes capital flow based on the “siphon effects” and “spillover effects” (Ma et al., 2020) [3]. On the one hand, capital concentration in central cities reduces its marginal return, leading to capital flow from central cities to peripheral ones, as the opening of HSR has promoted factors’ free mobility, that is, “spillover effects”. This mechanism can well explain the cross-regional investments of corporations. On the other hand, some scholars have found that the marginal returns of capital in central cities are not lower than those in peripheral ones as described in the neoclassical theory, which can attract capital from peripheral cities flowing to central ones, namely the “siphon effect”. This is because central cities have better infrastructure, larger market size, and higher production efficiency, generating the effect of increasing returns to scale. For example, Ma et al. (2020) [3] found that although the opening of HSR has promoted enterprises in central cities to invest in other places, in general, the capital flow direction is from peripheral cities to central ones, which means that “siphon effects” dominate. The third aspect is HSR and corporate behaviors. The existing literature mostly discusses the impact of the opening of HSR on the reactions of corporations from the perspective of information asymmetry. As the opening of HSR reduced transportation costs, it encouraged information intermediaries to communicate with enterprises face-to-face, which promoted the exchange of “soft information”, leading to reducing information asymmetry between corporations and external investors (Guo and Zhang, 2021) [28], reducing stock crash risks (Zhao et al., 2018) [29] and improving the accuracy of corporate earnings forecasts (Yang et al., 2019) [30].

2.1.3. Summary of Existing Literature

Through analyzing the relevant literature, the following problems are found. First, although the existing literature has discussed the motivations of corporate financialization in detail, no research has directly focused on the relationship between the opening of HSR and corporate financialization. Second, the existing literature analyzes the motivation of corporate financialization either from the “investment substitution” theory or the “reservoir” theory, which cannot uncover the essence of corporate financialization. Therefore, this paper innovatively links the opening of HSR with corporate financialization, which bridges the gap. Meanwhile, by combining both the “investment substitution” theory with the “reservoir” theory, we build a comprehensive theoretical framework to explore the possible mechanism between the opening of HSR and corporate financialization, which not only expands the research scope but also provides suggestive recommendations for policymakers to guide real entities to return to their main business.

2.2. Mechanism Analysis

The “investment substitution” theory believes that the yield difference between financial assets and physical assets is an important reason for corporate financialization (Demir, 2009) [12]. Since the financial crisis broke out in 2008, the profit margin of traditional manufacturing enterprises has continued to decline. On the contrary, the alternate prosperity of the capital market and the real estate market has changed the capital allocation mode of “production—exchange—reinvest—reproduction” of real enterprises and made manufacturing entities invest excessive funds into the financial industry, thus resulting in corporate financialization. There are also scholars who explain corporate financialization from the perspective of managerial myopia (Lazonick, 2011) [31]. It is believed that managements biased in favor of short-term benefits would lead them to abandon physical projects with a long investment cycle. Essentially, financialization from the perspective of managerial myopia has the same theoretical connotation as the “investment substitution” theory, which is that abnormal returns in the financial industry drive real enterprises to excessively allocate funds into financial assets under the premise of limited resources.

Fundamentally, under the “investment substitution” theory, it is necessary to encourage enterprises to increase physical investment so as to crowd out the financial assets, thus driving real enterprises to return to their main business. The following will discuss the inhibition of HSR on corporate financialization by the “crowding-out effects” of increasing physical investments. The theoretical framework is shown in Figure 1.

The opening of HSR can inhibit financialization by increasing innovation inputs. Firstly, since the opening of HSR promoted regional economic growth, it has attracted an influx of corporate innovative factors such as senior labor force, funds, technology, etc., which can enrich corporate innovative factor reserves for local enterprises and reduce the cost of innovative activities to a certain extent, thus lowering the threshold of corporate innovation. Furthermore, the industrial agglomeration caused by the opening of HSR has promoted knowledge integration among different industries and is conducive to generating new knowledge as well as causing knowledge spillovers (Bian et al., 2019) [32]. Corporates at the knowledge spillover center can well digest and absorb the new knowledge, which shortens the cycle of commercialization of innovation achievements and enables enterprises to obtain the abnormal returns brought by innovation activities faster. This can stimulate enterprises’ enthusiasm for innovation and increase innovation input. Secondly, the opening of HSR can alleviate the insufficient innovation investment caused by managerial myopia. The opening of HSR narrows the temporal–spatial distance so that it can reduce business travel time. This could easily stimulate external stakeholders to carry out field research, which enhances their supervision on corporates. For example, the opening of HSR has significantly reduced the number of absences of non-local independent directors from board meetings (Ye et al., 2020) [33]. As the representatives of shareholders, non-local independent directors improving their influence on the board of directors can effectively curb managerial myopia behaviors so that they may alleviate the insufficient investments in corporate innovation activities. Therefore, under the constraints of limited resources, increasing corporate innovation inputs will crowd out financial investments, which can prevent companies from being financialized under the “investment substitution” theory.

The opening of HSR can crowd out financial investments by increasing physical cross-regional investments. The opening of HSR breaks through the restriction of geographical distance and can encourage enterprises to expand outward (Wang et al., 2020) [5]. From the perspective of production, it has improved regional accessibility and broken the market segmentation, which enables companies to invest in the region with the lower price of production factors. Moreover, the release of railway freight capacity by the opening of HSR has reduced logistics costs, which also further improves its profit margin, thus stimulating enterprises to invest more in other places. From the perspective of making decisions, before the opening of HSR, the barrier of geographical distance caused information asymmetry between managers and potential cross-regional investment projects, which made it difficult for the managers to grasp the potential investment opportunities. Worse still, cross-regional investment also faces high communication costs and supervision costs (Zhuang et al., 2021) [6] without the establishment of HSR, which can severely erode investment returns, making cross-regional investment less attractive. Since the opening of HSR, enterprises can easily collect information about non-local investment projects and make accurate predictions, which can improve decisions’ rationality and reduce the probability of missing potential investment opportunities. In addition, the opening of HSR has also tied corporate headquarters and off-site branches much closer. It enhances controlling and supervising off-site subsidiaries, which can significantly reduce communication costs and supervision costs, thus improving the attractiveness of cross-regional physical investments. It is not difficult to see that the opening of HSR can crowd out financial investment by increasing cross-regional physical investment. Therefore, the following hypotheses are proposed:

H1. 

The opening of HSR can inhibit corporate financialization.

H2. 

The opening of HSR can inhibit corporate financialization by increasing innovation investments.

H3. 

The opening of HSR can inhibit corporate financialization by increasing cross-regional physical investments.

Another explanation of corporate financialization is the “reservoir” theory. It is derived from Keynes’ precautionary savings theory, which believes that the purpose of the allocation of financial assets is to enhance corporate liquidity. The credit system dominated by traditional commercial banks has serious credit discrimination against small and medium-sized enterprises (Tsai and Kellee, 2016) [34]. The lack of internal funds together with severe external financing constraints may make these enterprises miss potential investment opportunities. In this case, to cope with the liquidity crisis and smooth future investment, these enterprises have to increase their holdings of short-term liquid financial assets as their liquidity reserve, which makes them show such financialized features.

The financialization described by the “reservoir” theory is more like “proactive” behavior. Under internal and external financing constraints, enterprises can only reserve enough liquidity in advance by holding short-term financial assets to smooth future investments. Therefore, easing internal and external financing constraints and supplementing liquidity demand are effective means to drive enterprises to de-financialize. The following will discuss how the “liquidity replenishment effect” caused by the opening of HSR will restrain corporate financialization under the “reservoir” theory.

The opening of HSR can inhibit corporate financialization by increasing internal cash flows. First, the establishment of HSR has eliminated obstacles to the trans-regional flow of production factors. In order to attract advanced production factors, local authorities would introduce a series of preferential policies that can optimize the business environment of local enterprises, thus being conducive to the accumulation of internal cash flows. For example, from the perspective of tax policy, HSR has intensified the tax competition among local authorities (Pu et al., 2022) [35], which can significantly reduce the actual corporate tax burden, thus saving internal cash flows. Especially under the “official promotion tournament” in China, where local governments focus more attention on GDP growth, the political burden of local authorities intensifies tax competition among different regions, thus enhancing the effect of tax reduction on saving internal cash flows. Secondly, HSR services tie the enterprise with both upstream suppliers and downstream customers much closer, which can release internal cash flows from the supply chain. From the perspective of the relationship with suppliers, frequent business visits brought by HSR services may enhance the synergy effect between the two, which promotes the enterprise’s inventory management efficiency. This can release cash flows from the occupation of inventory and increase corporate liquidity. From the perspective of the relationship with clients, the business visits between enterprises and customers can enhance the commercial trust between the two, dispelling clients’ doubts about the quality of products, which can shorten the accounts receivable cycle (Chen and Liu, 2019) [36] and accelerate receivable turnover, thus releasing cash flows occupied by clients. It is not difficult to see that the opening of HSR has promoted the accumulation of internal cash flows, which supplies liquidity to enterprises so that it can effectively restrain corporate financialization under the “reservoir” theory.

The opening of HSR can inhibit corporate financialization by improving external financing capabilities. First, from the perspective of debt financing, the opening of HSR has generated economic agglomeration, which can attract huge potential clients for local banks. This can drive local banks to expand excessively. As the bank competition continues, commercial banks may lower the financing threshold and cut lending rates to attract more customers (Wu et al., 2021) [37], which objectively enhances companies’ financing capability and increases corporate liquidity. From the perspective of equity financing, the improvement of information exchange efficiency caused by HSR services can effectively enhance the information transparency of enterprises (Guo et al., 2021) [38] and can alleviate the information asymmetry between enterprises and investors so that it helps to reduce the cost of equity financing. Secondly, the opening of HSR provides enterprises with diversified financing channels. The opening of HSR has improved the convenience of field research for venture investors and face-to-face communication between investors and enterprises has increased. This can significantly promote the exchange of “soft information” including business environments, government-enterprise relationship as well as future development strategies, etc. Compared with hard information i.e., financial information, “soft information” can well reflect the enterprises potential (Guo et al., 2021) [38]. Through face-to-face communication, venture investors would improve their awareness of corporate soft information so that it can improve the probability of enterprises obtaining the venture capital (Long et al., 2019) [4]. Consequently, the opening of HSR has improved corporate liquidity by increasing internal cash flows and external financing capabilities, which can restrain corporate financialization under the “reservoir” theory. Therefore, the following hypotheses are proposed:

H4. 

The opening of HSR can inhibit corporate financialization by accumulating internal cash flows.

H5. 

The opening of HSR can inhibit corporate financialization by improving external financing capabilities.

3. Data and Methodology

3.1. Sample Selection and Data Source

This paper builds a time-varying DID model to empirically test the relationship between the opening of HSR and corporate financialization in China by matching the data of HSR at the city level with the data of non-financial A-share companies listed in the Shanghai and Shenzhen stock markets from 2007 to 2020. The financial data involved in this paper comes from the CSMAR database; the HSR data are from the official website of the National Railway Group Co., Ltd. (Beijing, China) and the city-level data are from the China Urban Statistical Yearbook. To improve the data quality, we cleaned the data in the following steps: First, eliminating samples in the financial and real estate industries; second, eliminating samples with special treatment (ST and * ST); third, eliminating samples in the process of IPO or delisting; fourth, eliminating samples with missing values. After the above treatment, 28,623 observations of 3528 enterprises were obtained. We also winsorized all variables at the 1% level.

3.2. Variable Interpretation

3.2.1. Dependent Variable

Financialization (Fin). According to the definition of corporate financialization, both the proportion of financial assets to total assets and the financial investment income can be used to describe corporate financialization (Du et al., 2019) [1]. However, considering that the financial investment income may include macrofactors and investors’ irrational noise, we selected the proportion of financial assets to total assets to represent corporate financialization. Specifically, trading financial assets, derivative financial assets, loans and advances, available-for-sale financial assets, held-to-maturity investments, and investment property were classified as financial assets. It should be noted that the two statement items of held-to-maturity investment and available-for-sale financial assets were deleted from the accounting standards in 2018 and later. Referring to the practice of Du and Zhou (2019) [18], in 2018 and later years, debt investment was used to replace held-to-maturity investment. Other debt investments together with other equity instrument investments were used to replace available-for-sale financial assets. In addition, since the proportion of financial assets will increase when non-financial assets fall unexpectedly, we also selected the absolute value of financial assets as an alternative indicator.

3.2.2. Independent Variable

High-speed railway opening (HSR). Based on the HSR data of 286 cities in China from 2007 to 2020, the dummy variable of HSR was constructed. Specifically, if the city where the enterprise is located established a high-speed railway station during the sample period, the value of HSR is 1 in the current year and later years; otherwise, it is 0. According to the practice of Bian et al. (2019) [32], if more than one high-speed railway station was built in the city, the year when HSR was opened earlier shall be taken as the opening time. In addition, we also introduced the number of HSR stations as an alternative independent variable to further improve the robustness of the conclusion.

3.2.3. Control Variables

To overcome the impact of missing variables as much as possible, a series of control variables were introduced including firm size (Size), return on assets (Roa), leverage (Lev), capital intensity (Den), ownership concentration (EqDen), company growth (Tobinq), board size (Board), firm age (Age) and duality (Dual) (Du et al., 2019; Du and Zhou, 2019) [1,18]. The detailed definition of all variables is shown in

Table 1. Variable definitions.

Variable Type	Variable Symbol	Variable Definition
Dependent Variables	Fin1	Total Financial Assets/Total Assets
	Fin2	The Nature Log of Total Financial Assets
Independent	HSR	Cities with HSR opened is 1; otherwise, is 0
Variables	Station	The nature log of the numer of HSR stations
	Size	The nature log of total assets
	Roa	Net profit/total assets
	Lev	Total debts/total assets
Control	Den	Total assets/total income
Variables	Eqden	Shareholding ratio of the larggest shareholder
	Tobinq	Market value/total asset
	Board	The nature of the number of board directors
	Age	The nature log of firm age
	Dual	Chairman and CEO are the same person is 1, otherwise is 0

.

3.3. Benchmark Regression Model Specification

This paper takes the cities with HSR opened as the treated group and cities without as the control group and builds a time-varying DID model to test the relationship between HSR and corporate financialization. The model settings are as follows:

$$Fi{n}_{i,t}={\alpha }_0+a_{1}trea{t}_i*pos{t}_t+{{\displaystyle \sum }}_j{a}_{j}C{V}_{j,i,t}+{\lambda }_i+{\delta }_d+{\phi }_t+{\epsilon }_{i,d,t}$$

(1)

where i refers to companies; d stands for industries; j represents the number of control variables; c refers to the cities where firms are located; t represents time. Fin_i,t refers to the degree of financialization of company i in year t, which is represented by two indicators: Fin1 and Fin2. The independent variable is HSR_c,t which is represented by $trea{t}_i* pos{t}_t$, where treat_i is a treatment group dummy variable that takes the value of 1 (treatment group) when company i’s city has opened HSR and 0 (control group) if company i’s city has not opened HSR. post _t is a treatment period dummy variable that takes the value of 1 when company i enters the treatment period and 0 otherwise. $trea{t}_i* pos{t}_t$ is used to capture the difference between the change before and after the treatment group and control group, i.e., the treatment effect. CV is the control variable set. To ensure the accuracy of the results, we also controlled the fixed effects at the firm, industry, and year levels, which are represented by λ_I, δ_d φ_t, respectively. In addition, we adopted statistics adjusted by the cluster robust standard error in all regressions and clustered the standard error of the estimated coefficient to firm level.

4. Empirical Results

4.1. Descriptive Statistics

Table 1 reports the summary statistics of the variables. Specifically, the average value of corporate financialization (Fin1) is 0.0366, which means the average degree of corporate financialization among sample firms is 3.66%. In addition, some sample companies suffer severe financialization since the maximum of corporate financialization is 0.4113. Furthermore, the average value of HSR is 0.7878, indicating that 78.78% of the cities have opened HSR. In terms of control variables, the mean values of Size, Roa, Lev and Den are 22.0325, 0.0406, 0.4187, 2.4390, respectively, which are almost consistent with the relevant studies. The rest of control variables are also within a reasonable range; for more details, see

Table 2. Descriptive statistics.

Variable	Obs	Mean	Median	Std. Dev.	Minimum	Maximum
Fin1	28,623	0.0366	0.0063	0.0719	0.0000	0.4113
Fin2	28,623	0.8045	0.1419	1.5724	0.0000	10.5296
HSR	28,623	0.7878	1.0000	0.4089	0.0000	1.0000
Station	28,623	1.3364	1.6094	0.8684	0.0000	3.2189
Size	28,623	22.0325	21.8545	1.2942	19.5657	27.2028
Roa	28,623	0.0406	0.0397	0.0587	−0.2282	0.1920
Lev	28,623	0.4187	0.4115	0.2061	0.0506	0.9310
Den	28,623	2.4390	1.8310	2.7420	0.3881	37.0932
Eqden	28,623	0.3496	0.3299	0.1492	0.0893	0.7482
Tobinq	28,623	2.0673	1.6486	1.3032	0.8783	8.3525
Board	28,623	2.2540	2.3026	0.1784	1.7918	2.7726
Age	28,623	2.7990	2.8332	0.3680	1.6094	3.4657
Dual	28,623	0.2679	0.0000	0.4429	0.0000	1.0000

.

4.2. Benchmark Regression Results

Table 3. Benchmark regression results.

	Fin1		Fin2
	(1)	(2)	(3)	(4)
HSR	−0.0049 *** (−3.33)	−0.0045 *** (−3.10)	−0.1064 *** (−3.30)	−0.0981 *** (−3.07)
Size		−0.0033 ** (−2.07)		−0.0363 (−1.04)
Roa		−0.0342 *** (−3.42)		−0.7106 *** (−3.33)
Lev		−0.0289 *** (−5.00)		−0.6539 *** (−5.24)
Den		0.0023 *** (4.46)		0.0500 *** (4.52)
Eqden		0.0072 (0.70)		0.2075 (0.92)
Tobinq		0.0026 *** (4.14)		0.0529 *** (3.96)
Board		−0.0019 (−0.37)		−0.0382 (−0.34)
Age		0.0391 *** (4.56)		0.8146 *** (4.34)
Dual		−0.0004 (−0.24)		−0.0036 (−0.10)
Constant	0.0404 *** (34.90)	0.0077 (0.19)	0.8883 *** (34.94)	−0.5116 (−0.57)
Control Variables	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Adjusted-R2	0.5567	0.5646	0.5610	0.5681
Obs	28,623	28,623	28,623	28,623

Notes: t statistics adjusted by clustered standard error are in parentheses; *** and ** are significant at the level of 1% and 5%, respectively.

reports the benchmark regression result. Colum (1) and (2) represent the regress results of Fin1, while Colum (3) and (4) report the results with Fin2. Specifically, the estimated coefficients of HSR in column (1) and column (3) are −0.0049 and −0.1064 respectively, without adding the control variables, which are significant at the level of 1%. It can be preliminarily judged that the opening of HSR can inhibit corporate financialization. To prevent other factors from interfering with the regression result, a series of control variables are added in columns (2) and (4). The estimated coefficient is slightly lower than that in columns (1) and (3), but it is still significant at the level of 1%, which suggests that some of the impact of HSR on corporate financialization is absorbed by the control variables, but HSR’s effect is still statistically significant. Therefore, it can be concluded that the opening of HSR can inhibit corporate financialization, which supports hypothesis H1.

4.3. Endogenous Treatment and Robustness Test

4.3.1. Parallel Trend Test

We conducted the parallel trend test to verify whether the time-varying DID model meets the parallel trend hypothesis. We referred to the method of Zhu et al. (2019) [39]. The model is set as follows:

$$Fi{n}_{i,t}={\gamma }_1+{{\displaystyle \sum }}_{-7}^6{\gamma }_{2}HS{R}_{c,t}+{{\displaystyle \sum }}_j{\gamma }_{j}CV+{\lambda }_i+{\delta }_d+{\phi }_t+{\epsilon }_{i,d,t}$$

(2)

where ${\displaystyle \sum }_{-7}^6{\gamma }_{2}HS{R}_{c,t}$ refers to the degree of corporate financialization in early years (7–1 years before HSR is opened), the current year that HSR is opened, and later years (1–6 years after HSR is opened). The other variables are the same as model (1).

Table 4. Parallel trend test.

	Fin1	Fin2
	(1)	(2)
HSR (−7)	−0.0070 (−0.78)	−0.1677 (−0.85)
HSR (−6)	−0.0071 (−0.90)	−0.1665 (−0.96)
HSR (−5)	−0.0073 (−0.99)	−0.1693 (−1.06)
HSR (−4)	−0.0056 (−0.82)	−0.1357 (−0.92)
HSR (−3)	−0.0070 (−1.13)	−0.1661 (−1.22)
HSR (−2)	−0.0067 (−1.18)	−0.1584 (−1.28)
HSR (−1)	−0.0071 (−1.39)	−0.1637 (−1.46)
HSR (0)	−0.0082 * (−1.78)	−0.1874 * (−1.86)
HSR (+1)	−0.0101 ** (−2.47)	−0.2286 ** (−2.55)
HSR (+2)	−0.0077 ** (−2.13)	−0.1765 ** (−2.24)
HSR (+3)	−0.0081 ** (−2.58)	−0.1841 *** (−2.69)
HSR (+4)	−0.0090 *** (−3.47)	−0.2036 *** (−3.57)
HSR (+5)	−0.0074 *** (−3.52)	−0.1663 *** (−3.63)
HSR (+6)	−0.0083 *** (−5.29)	−0.1833 *** (−5.37)
Constant	0.0043 (0.10)	−0.5765 (−0.64)
Control Variables	YES	YES
Firm FE	YES	YES
Industry FE	YES	YES
Year FE	YES	YES
Adjusted-R2	0.5650	0.5682
Obs	28,623	28,623

***, **, * are significant at the level of 1%, 5% and 10% respectively.

reports the result of the parallel trend test. Specifically, the estimated coefficients of early years (7–1 years before the opening of HSR) in column (1) and column (2) are not significant and the estimated coefficients of HSR are significantly negative in the later years, indicating that there is no significant difference between the treated group and the control group before the opening of HSR, which meets the parallel trend hypothesis. These results prove that the estimated results in the benchmark regression are unbiased.

4.3.2. Placebo Test

In the benchmark regression, this paper draws the conclusion that the opening of HSR can inhibit corporate financialization. It is undeniable that although the fixed effects of firms, industries, and years are controlled in the benchmark regression, they cannot eliminate the unobservable systematic differences that affect corporate financialization between the treated group and the control group, that is, the inhibition of the opening of HSR on corporate financialization is likely caused by those unobservable systematic differences, rather than the opening of HSR. In order to solve the above problem, referring to the practice of Ma et al. (2020) [3], we randomly selected a number of cities from the cities with HSR opened and randomly generated the opening time of HSR so as to form a randomly generated treatment group, then re-ran the regression in the model (1) based on the randomly generated dataset (HSR_Random), and repeated the regression 500 times by the Monte Carlo model, and finally mapped the distribution of the estimated coefficient. As shown in Figure 2, for both Fin1 and Fin2, the estimation coefficients of HSR_Random are all distributed around 0, while the real benchmark regression estimation coefficients are obviously out of the distribution of HSR_Random; there are only 2.6% and 3.4% probabilities that the estimated coefficients are in the distribution of HSR_Random. In short, the result passed the placebo test, implying that the inhibition of corporate financialization by the opening of HSR is not caused by accidental factors, and the conclusion still holds.

4.3.3. Instrumental Variable Estimation Approach

The location of the high-speed railway station is not randomly selected, but will be affected by the local population size, economic development potential, and political considerations. These factors are difficult to control and are likely to have an impact on corporate financialization, resulting in serious endogenous problems. Therefore, it is necessary to employ the instrumental variable estimation approach to further deal with endogenous problems.

This paper selected two instrument variables. First, referring to the practice of Faber (2014) [40], a minimum spanning tree constructed based on geographical information was selected. The rationality of the variable is that, on the one hand, geographical information is an important factor affecting the construction of HSR, which meets the relevance requirements. On the other hand, local geographic information is difficult to directly affect the corporate financialization so that it meets the exogenous requirements. Second, with reference to the methods of Liu and Li (2017) [41], the total railway passenger transport volume of sample cities in China in 1990 was selected as the second instrumental variable. The historical passenger traffic volume reflects the development level of urban transportation facilities, from which the possibility of building HSR in the local area can be deduced, thus meeting the relevance requirements. The historical passenger volume cannot directly affect corporate financialization and meets exogenous requirements. It should be noted that the minimum spanning tree and historical passenger volume are both cross-sectional data, which do not change with time and cannot directly match with panel data. In this regard, we constructed the interaction item as the proxies by multiplying the minimum spanning tree and 1990 historical passenger volume with the year-dummy variable and then introduced them to the model (2SLS).

The results of instrumental variable estimation are shown in

Table 5. Instrumental variable estimation.

	IV_Tree		IV_Hist
	Fin1	Fin2	Fin1	Fin2
	(1)	(2)	(3)	(4)
HSR	−0.0598 ** (−2.37)	−1.3638 ** (−2.44)	−0.0478 *** (−2.86)	−1.0772 *** (−2.95)
Control Variables	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Obs	24,879	24,879	25,337	25,337
Kleibergen–Paap rk F	40.9290		91.3620
IV First Stage	−0.1443 ***		−0.0087 ***

*** and ** are significant at the level of 1% and 5% respectively.

. Firstly, the values of Kleibergen–Paap rk F statistics of the two instrumental variables are 32.656 and 91.362, respectively, which are greater than the Stock and Yogo critical values of all levels, indicating that the hypothesis of weak instrumental variables can be rejected. The estimated coefficients in the first stage of the two instrumental variables are significant at the level of 1%, and the sign of two estimated coefficients is consisted with the expectations. In short, the above statistical results proved that the instrumental variables are valid. Secondly, in columns (1)–(4), the estimated coefficient of HSR is significantly higher than that of the benchmark regression (−0.0045 and −0.0981), which suggests that there are indeed endogenous problems in the model, and if it is not handled, the estimated results will be misinterpreted. Most importantly, in columns (1)–(4), the estimated coefficients of HSR are significantly negative at the level of 1%, implying that after considering the above endogenous problems, the regression results of the instrumental variables estimation still support the basic conclusion, that is, the opening of HSR can significantly restrain corporate financialization, thus confirming that the basic conclusion is robust.

4.3.4. PSM-DID

To further eliminate the interference of sample self-selection, a control group was constructed by employing the propensity score matching method (PSM). Specifically, referring to the practice of Ji and Yang (2020) [26], we selected the firm size (Size), profit before interest and tax (Income), asset-liability ratio (Lev), capital intensity (Den), financing constraint index (KZ), and equity concentration (Eqden) as the matching covariates to carry out the k-nearest neighbor 1:1 sampling matching in the caliper (caliper = 0.05). On this basis, we re-examined the impact of the opening of HSR on corporate financialization.

To ensure the reliability of the propensity score matching results, the balance test of covariates was also carried out, and results are reported in Appendix A. Specifically, after matching, the deviation of covariates between the treated group and the control group decreased to less than 1%; the LR statistics decreased from 1476.42 to 7.92, suggesting that there were no significant systematic differences between the treated group and the control group, which meets the balance hypothesis. The results of common support in Figure 3 show that after the correction of propensity score matching, the distribution of the propensity score of the treated group and the control group tended to coincide, which meets the common support hypothesis.

Table 6. PSM-DID.

	Fin1	Fin2
	(1)	(2)
HSR	−0.0040 *** (−2.69)	−0.0871 *** (−2.63)
Constant	0.0225 (0.52)	−0.1505 (−0.16)
Control Variables	YES	YES
Firm FE	YES	YES
Industry FE	YES	YES
Year FE	YES	YES
Adjusted-R2	0.5700	0.5740
Obs	25,106	25,106

*** is significant at the level of 1%.

reports results of PSM-DID. First, in columns (1) and (2), the estimated coefficients of HSR are −0.0040 and −0.0871, respectively, which are significantly lower than the results of the benchmark regression (−0.0045 and −0.0981, respectively), indicating that factors that cannot be controlled in the model indeed affected the benchmark results, and if they are not controlled, the HSR coefficients in benchmark regression will be overestimated. Second, after the correction of the propensity score matching, the estimated coefficient of HSR is significantly negative at the level of 1% for both Fin1 and Fin2, which implies that the opening of HSR can indeed inhibit corporate financialization after considering the sample self-selection problem, thus further proving that the basic conclusion is robust.

4.3.5. Other Robustness Tests

First, considering the interference of the measurement error of the independent variable, we replaced the independent variable HSR with Station, that is, the number of high-speed railway stations owned by the city where the enterprise office is located, and retested. The results are shown in columns (1) and (2) of

Table 7. Other robustness tests.

	Replacing Independent Variable		Shortening Sample Period		Eliminating Samples in Central Cities
	Fin1	Fin2	Fin1	Fin2	Fin1	Fin2
	(1)	(2)	(3)	(4)	(5)	(6)
Station	−0.0020 * (−1.86)	−0.0412 * (−1.75)
HSR			−0.0036 * (−1.75)	−0.0807 * (−1.84)	−0.0048 ** (−2.55)	−0.0994 ** (−2.40)
Constant	0.0065 (0.16)	−0.5442 (−0.61)	0.1821 ** (2.23)	3.3653 * (1.92)	0.0185 (0.35)	−0.0926 (−0.08)
Control Variables	YES	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
Adjusted-R2	0.5645	0.5680	0.7487	0.7523	0.5538	0.5546
Obs	28,623	28,623	7712	7712	14,548	14,548

**, * are significant at the level of 5% and 10% respectively.

. The estimated coefficient of Station is negative at least at the level of 10%. Second, since the sample period includes special periods such as the financial crisis in 2008 and the Chinese stock disaster in 2015 events such as those will have a systematic impact on the financial industry, thus interfering with the basic conclusions. In this regard, we shortened the window period and took samples from 2012–2014 to retest. The results are shown in columns (3) and (4) of Table 7. The estimated coefficient of HSR is at least negative at the level of 10%, which supports the basic conclusion. Third, to further control the non-randomness of the location of high-speed railway, we eliminated samples in central cities (including 27 provincial capital cities and 4 municipalities directly under the Central Government) with the reference to the practice of Zhang et al. (2018). The results are shown in columns (5) and (6) of Table 7. The estimated coefficients of HSR are significantly negative at the level of 5%. Therefore, the basic conclusion is strongly robust after considering the above robustness checks.

4.4. Mechanism Analysis

According to the previous analysis, this paper theoretically discusses how the opening of HSR might inhibit corporate financialization from the perspective of both “investment substitution” theory and “reservoir” theory. This part will empirically test Hypotheses H2–H5 with the help of the mediating effect model. The specific model settings are as follows:

$$Fi{n}_{i,t}={\alpha }_0+a_{1}HS{R}_{c,t}+{{\displaystyle \sum }}_j{a}_{j}C{V}_{j,i,t}+{\lambda }_i+{\delta }_d+{\phi }_t+{\epsilon }_{i,d,t}$$

(3)

$$Me{d}_{i,t}={\beta }_0+{\beta }_{1}HS{R}_{c,t}+{{\displaystyle \sum }}_j{\beta }_{j}C{V}_{j,i,t}+{\lambda }_i+{\delta }_d+{\phi }_t+{\epsilon }_{i,d,t}$$

(4)

$$Fi{n}_{i,t}={\eta }_0+{\eta }_{1}HS{R}_{c,t}+{\eta }_{2}Me{d}_{i,t}+{{\displaystyle \sum }}_j{\eta }_{j}C{V}_{j,i,t}+{\lambda }_i+{\delta }_d+{\phi }_t+{\epsilon }_{i,d,t}$$

(5)

where model (3) is the same as the benchmark regression, and model (4) is used to test the influence of the independent variables on the mediator variables; model (5) includes both independent variables and mediator variables. If α₁, β₁, η₁ and η₂ are significant at the same time, and β₁η₂ and η₁ have the same sign, it proves that the mediating effect exists.

In addition, this paper selected innovation investment (Inno) and physical investment (Invest) to test the mechanism of HSR opening to corporate financialization under the “investment substitution” theory. Referring to the practice of Du et al. (2019), the natural logarithm of the total R&D investment was used as an indicator to measure the innovation investment. Referring to the research of Du and Zhou (2019) [18], the net amount of fixed assets, the net amount of construction in progress, and the total amount of construction materials were defined as the physical investment, and they were summed up and logarithmized to measure the physical investment. At the same time, we selected the nature logarithm of internal capital flow (Cfo) as the proxy of internal capital, and selected the external financing constrains index (KZ) as the proxy of corporate financing capabilities to test the mechanism of HSR opening to corporate financialization under the “reservoir” theory.

Table 8. Mechanism analysis: innovation investment.

	Fin1	Fin2	Inno	Fin1	Fin2
	(1)	(2)	(3)	(4)	(5)
HSR	−0.0062 *** (−3.63)	−0.1338 *** (−3.50)	0.0725 ** (2.06)	−0.0061 *** (−3.51)	−0.1300 *** (−3.38)
Inno				−0.0025 *** (−2.84)	−0.0543 *** (−2.82)
Constant	−0.0308 (−0.70)	−1.1954 (−1.23)	1.5660 ** (2.35)	−0.0268 (−0.61)	−1.1103 (−1.15)
Control Variables	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
Sobel (Z)				2.8200 ***	2.8180 ***
Proportion of Mediating Effects				2.92%	2.94%
Adjusted-R2	0.5250	0.5316	0.8636	0.5254	0.5320
Obs	21,828	21,828	21,828	21,828	21,828

*** and ** are significant at the level of 1% and 5% respectively.

reports the mediating effect test results of the innovation investment. Column (1) and (2) reports the result of model (3), column (3) presents the result of model (4), and column (5) and (6) document the result of model (5). As the estimated coefficients are significant in columns (1) and (2), the mediating effects of the innovation investment are further considered in column (3)–(5). The results in column (3) show that the estimated coefficient of HSR is 0.0725 and is significant at the level of 1%, implying that the opening of HSR can significantly increase corporate innovation investment. Moreover, the estimated coefficients of HSR and Inno are significantly negative in column (4) and (5); the absolute value of HSR’s estimation coefficient has decreased compared with columns (1) and (2) and β₁η₂ and η₁ have the same sign; accordingly, it can be preliminarily determined that the mediating effect of innovation investment exists. The results of the Sobel test show that Sobel’s Z statistics in columns (4) and (5) are both at the level of 1%, proving that the the mediating effect of innovation investment does exist. The proportion of the mediating effect is 2.92% ([0.0725 × (−0.0025)]/−0.0062) and 2.94% ([0.0725 × (−0.0543)]/−0.1338)), respectively. Therefore, the above results show that the opening of HSR can indeed inhibit corporate financialization by increasing innovation investment, that is, the path of “HSR → (increase) innovation investment → (inhibit) corporate financialization” is established, thus supporting hypothesis H2.

Table 9. Mechanism analysis: physical investment.

	Fin1	Fin2	Invest	Fin1	Fin2
	(1)	(2)	(3)	(4)	(5)
HSR	−0.0045 *** (−3.08)	−0.0976 *** (−3.06)	0.0371 ** (2.25)	−0.0041 *** (−2.85)	0.0900 *** (−2.93)
Invest				−0.0092 *** (−5.94)	−0.2047 *** (−6.16)
Constant	−0.0090 (0.22)	−0.4861 (−0.54)	−0.8146 * (−1.71)	0.0015 (0.04)	−0.6529 (−0.73)
Control Variables	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
Sobel (Z)				−3.2160 ***	−3.2190 ***
Proportion of Mediating Effects				7.58%	7.78%
Adjuted-R2	0.5647	0.5681	0.9348	0.5676	0.5712
Obs	28,617	28,617	28,617	28,617	28,617

***, **, * are significant at the level of 1%, 5% and 10% respectively.

reports the results of the mediating effect tests of physical investment. Since estimated coefficients of HSR in column (1) and (2) are significant, further analysis was conducted. The results in column (3) show that the estimated coefficient of HSR is 0.0371 and is significant at the level of 1%, indicating that the opening of HSR can increase the physical investments. The results in column (4) and (5) show that the estimated coefficients of HSR and physical Investment are significantly negative at the level of 1%, and the absolute value of the estimated coefficients of HSR is significantly lower than that of (1) and (2), which can preliminarily determine that the mediating effect of physical investment is tenable. The results of Sobel test show that the Z statistics in columns (4) and (5) are significant at the 1% level, and the proportion of the mediating effect is 7.58 ([0.0371 × (−0.0092)]/−0.0045) and 7.78% ([0.0371 × (−0.2047)]/−0.0976), which once again confirms that the physical investment plays a partial mediator between the opening of HSR and corporate financialization, that is, the transmission path of “HSR → (increase) physical investment → (inhibit) corporate financialization” is established, which supports hypothesis H3.

Table 10. Mechanism analysis: internal cash flows.

	Fin1	Fin2	Cfo	Fin1	Fin2
	(1)	(2)	(3)	(4)	(5)
HSR	−0.0044 *** (−2.81)	−0.0953 *** (−2.77)	0.0501 ** (1.99)	−0.0043 *** (−2.76)	−0.0936 *** (−3.02)
Cfo				−0.0016 *** (−3.45)	−0.0348 *** (−3.44)
Constant	−0.0184 (0.43)	−1.0005 (−1.07)	−1.7124 *** (−3.40)	−0.0211 (−0.50)	−1.0601 (1.13)
Control Variables	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
Sobel (Z)				−1.8480 *	−1.8450 *
Proportion of Mediating Effects				1.82%1	1.83%
Adjusted-R2	0.5485	0.5510	0.7619	0.5488	0.5513
Obs	22,881	22,881	22,881	22,881	22,881

***, **, * are significant at the level of 1%, 5% and 10% respectively.

reports the mediating effect test results of internal cash flows. Since the estimated coefficients of HSR in column (1) and (2) are significant, following analysis were considered. The results in column (3) reflect the influence of independent variables on mediator variables (Cfo), the estimated coefficient of Cfo is 0.0502, which is significant at the level of 5%, implying that the opening of HSR can significantly increase corporate internal capital flows. The estimated coefficients of HSR and Cfo in column (4) and (5) are significantly negative at the level of 1%, and the absolute value of the HSR estimated coefficients is lower than that of columns (1) and (2); β₁η₂ has the same sign as the η₁. In short, it can be preliminarily concluded that a partial mediating effect of internal cash flow exists. The Sobel test results show that the Z statistic in the (4) and (5) columns is −1.85 and is significant on the 10% level, indicating that the mediating effects of internal cash flows do exist. The proportion of the mediating effects is 1.83% ([0.0502 × (−0.0016)]/−0.0044) ([0.0502 × (−0.0348)]/−0.0953). Therefore, the above results show that the transmission mechanism of increasing internal cash flows to inhibit corporate financialization is tenable, that is, the HSR → (increasing) internal cash flows → (inhibiting) corporate financialization, which supports hypothesis H4.

Table 11. Mechanism analysis: external financing capability.

	Fin1	Fin2	KZ	Fin1	Fin2
	(1)	(2)	(3)	(4)	(5)
HSR	−0.0045 *** (−3.10)	−0.0981 *** (−3.07)	−0.0796 ** (−2.20)	−0.0044 *** (−3.04)	−0.0960 *** (−3.02)
KZ				0.0013 *** (4.15)	0.0269 *** (3.94)
Constant	0.0077 (0.19)	−0.5120 (−0.57)	−2.7649 *** (−3.64)	0.0113 (0.28)	0.4376 (−0.49)
Control Variables	YES	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES
Sobel (Z)				−2.2580 **	−2.2330 **
Proportion of Mediating Effects				2.30%	2.18%
Adjusted-R2	0.5646	0.5681	0.7254	0.5651	0.5685
Obs	28,622	28,622	28,622	28,622	28,622

*** and ** are significant at the level of 1% and 5% respectively.

reports results of the mediating effects of external financing capacity. The estimated coefficients of HSR in columns (1) and (2) are significantly negative so that the following analysis was considered. The estimated coefficient of HSR in column (3) is −0.0796, which is significant at the level of 5%, suggesting that the opening of HSR can improve the external financing capability of enterprises. The estimated coefficients of HSR in column (4) and (5) are significantly negative at the level of 1%, and the absolute value of the estimated coefficient in columns (1) and (2) has decreased; the estimated coefficient of KZ is significantly positive at the level of 1%, and β1η2 and η1 have the same sign. Therefore, it can be preliminarily determined that the external financing capacity plays a partial mediating effect between the opening of HSR and corporate financialization. The subsequent Sobel test results show that the Z statistics are 2.26 and 2.23, respectively, which are significant at the level of 5%. It has been confirmed that the external financing capacity plays a partial mediating effect, and the proportion of the mediating effect for Fin1 and Fin2 is 2.30% (−0.0796 × 0.0013/−0.0045) and 2.19% (−0.0796 × 0.0269/−0.0981), respectively. In conclusion, the above results show that the opening of HSR can indeed inhibit corporate financialization by improving the external financing capacity, that is, the path of “HSR → (improving) external financing capacity → (inhibiting) corporate financialization” exists, which provides empirical evidence for hypothesis H5.

4.5. Heterogeneity Analysis

4.5.1. Considering the Difference in Firm Size

The firm size will have a heterogeneous impact on the relationship between the opening of HSR and corporate financialization. On the one hand, large enterprises are the backbone of local economic development and will attract more attentions from local authorities. After the opening of HSR, local government officials will face more severe inter-city competition, and they will face greater political pressure to improve local economy. Therefore, they are more likely to enact preferential policy to support large enterprises in terms of credit, tax, subsidies, etc., aiming to help local firms win the competition. This can objectively promote the accumulation of internal cash flows since the policies create a business-friendly environment, which can restrain corporate financialization by “liquidity replenishment effects”.

The sample was divided into two groups by the median of enterprise total assets in the same industry. Specifically, if the sample’s total asset was bigger than the median of total assets in the same industry, we denoted them as the large size; otherwise, we denote them as small size. The results are reported in

Table 12. Heterogeneity analysis: firm size.

	Fin1		Fin2
	Large Size	Small Size	Large Size	Small Size
	(1)	(2)	(3)	(4)
HSR	−0.0048 ** (−2.34)	−0.0021 (−1.05)	−0.1070 ** (−2.26)	−0.0430 (−1.03)
Constant	−0.0013 (−0.02)	−0.0268 (−0.33)	−0.7424 (−0.51)	−1.1573 (−0.68)
Control Variables	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Adjusted-R2	0.6404	0.5850	0.6405	0.5886
Obs	14,169	14,093	14,169	14,093
p-Value	0.016 **		0.012 **

** is significant at the level of 5%.

. For both Fin1 and Fin2, the estimated coefficients of independent variable (HSR) are significant in the sample group with a large firm size but not significant in sample group without a large size. The above results show that the opening of HSR has a heterogeneous impact on corporate financialization, that is, the restrain effects of HSR are more significant in samples with a large firm size. We also employed the bootstrap method (500 times) to test the inter-group coefficient to further confirm the heterogeneity impact of the firm size. The results show that the p-values are 0.016 and 0.012, respectively, which are significant at the level of 5%, indicating that there are indeed differences in the impact of the opening of HSR on corporate financialization between the two groups, which supports the above assumption.

4.5.2. Considering the Difference in Industry Competition

The degree of industry competition has a heterogeneous impact on the relationship between the opening of HSR and corporate financialization. As a competitive mechanism, industry competition can influence the strategic decisions of enterprises through predatory pricing and entry threats. In the highly competitive industry, the market share of enterprises is decreasing and may even be crowd out. The opening of HSR can encourage enterprises to cultivate their core competitiveness by increasing corporate innovation investment. That is, the opening of HSR has a stronger marginal effect in highly competitive industries.

To verify the above assumption, the Hufendar index was used as the proxy of industry competition. The higher the Hufendar index, the lower the level of competition in the industry. Specifically, the sample was divided into two sample groups, namely, fierce competition and less competitive based on the median of the Hufendar index in the same industries. The results are reported in

Table 13. Heterogeneity analysis: industry competition.

	Fin1		Fin2
	Fierce	Not So Fierce	Fierce	Not So Fierce
	(1)	(2)	(3)	(4)
HSR	−0.0060 *** (−3.20)	−0.0013 (−0.62)	−0.1291 *** (−3.20)	−0.0272 (−0.57)
Constant	0.0542 (1.06)	0.0097 (0.18)	0.5765 (0.53)	−0.5197 (−0.43)
Control Variables	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Adjusted-R2	0.6332	0.6081	0.6373	0.6116
Obs	14,286	13,887	14,286	13,887
p-Value	0.000 ***		0.000 ***

*** is significant at the level of 1%.

. In general, the estimated coefficients of HSR opening were significant in the sample group with fierce industry competition but not significant in sample group without fierce industry competition. The p-value of intro-group difference test confirmed the existence of the heterogeneity of industry competition, which is consistent with expectations.

4.5.3. Considering the Difference in Initial Traffic Resource

The initial traffic resource of the city can have a heterogeneous impact on the opening of HSR and corporate financialization. Cities with a high initial traffic resource represent the improvement of local infrastructure, which can significantly reduce the traffic costs, and encourage information intermediaries to conduct field visits to enterprises. The “soft information” obtained by the information intermediary through field investigation can reduce the information asymmetry between investors and enterprises and improve the possibility of enterprises obtaining external financing, which can inhibit corporate financialization under the “reservoir” theory. In addition, compared with cities with low initial traffic resources, cities with high initial traffic resources can transform the advantages of traffic resources into innovative advantages, that is, abundant traffic resources promote the cross-regional flow of innovative factors, thus generating the agglomeration effect. The continuous integration of innovation factors promotes the creation of new knowledge, which can significantly reduce the innovation cost of local enterprises and encourage enterprises to increase innovation investment, thus playing the crowding-out role of innovation investment on financial investment and restraining the financialization under the “investment substitution” theory.

Based on this, this paper considers whether sample cities have an airport as the proxy of the initial traffic resource. If sample cities have an airport, denote them as the sample group with high initial traffic resource; otherwise, denote them as the sample group with poor initial traffic resource. The results are reported in

Table 14. Heterogeneity analysis: initial traffic resource.

	Fin1		Fin2
	High Initial Traffic Resource	Poor Initial Traffic Resource	High Initial Traffic Resource	Poor Initial Traffic Resource
	(1)	(2)	(3)	(4)
HSR	−0.0055 *** (−3.21)	0.0040 (1.49)	−0.1226 *** (−3.27)	0.0950 (1.62)
Constant	0.0005 (0.01)	−0.0410 (−0.51)	−0.7050 (−0.66)	−1.4710 (−0.84)
Control Variables	YES	YES	YES	YES
Firm FE	YES	YES	YES	YES
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Adjusted-R2	0.5747	0.5462	0.5787	0.5456
Obs	22,178	6416	22,178	6416
p-Value	0.000 ***		0.000 ***

*** is significant at the level of 1%.

. Specifically, for both Fin1 and Fin2, the impact of HSR opening on corporate financialization is more significant in the sample group with high initial traffic resource. In addition, the p-value of inter-group coefficient is significant at the level of 1%, which once again confirmed that the heterogeneity of the initial traffic resource does exist. In short, this conclusion is consistent with above expectations.

5. Conclusions

5.1. Discussions

The “shifting from real to virtual” of the real economy has become a characteristic fact of China’s economic development. At the microlevel, the corporate financialization leads to enterprises deviating from their main businesses and tending to hollow out the manufacturing industry, which not only impedes the high-quality development of enterprises but also exacerbates the instability of the financial system. Under this context, this paper considers the opening of HSR as a quasi-natural experiment and selects the non-financial listed companies in the Shanghai and Shenzhen A-share stock market from 2007 to 2020 as samples to systematically examine the relationship between the opening of HSR and corporate financialization. The study found that (1) The opening of HSR has a significant negative impact on corporate financialization, that is, the opening of HSR can significantly inhibit corporate financializaton. The conclusion still holds after a series of robustness tests. (2) The mechanism analysis shows that the opening of HSR can inhibit corporate financialization by increasing investment in innovation and enhancing physical investment, which are crowding-out effects, and by promoting the accumulation of internal capital flows and improving external financing capability, which are “liquidity replenishment effects”. (3) The inhibition of the opening of HSR on corporate financialization is more significant in large enterprises, competitive industries, and cities with high initial traffic resources.

5.2. Implications

According to the research results, the following policy recommendations are put forward: First, local authorities should fully digest and absorb the spirit of the report of the 20th CPC National Congress and put more attention on the real economy and promote the fit between manufacturing and financial industries to alleviate the structural imbalances of the national economy. Second, governments should persist in the strategy of strengthening the country through transportation and continue to increase the construction of transportation infrastructure in order to stimulate the catalytic role of transportation infrastructure in regional economic development. Third, governments should further eliminate the institutional barriers to the trans-regional flow of production factors and release the HSR’s potential for regional economic development. Fourth, non-central cities should clarify their industrial positioning and give full play to their comparative advantages in order to create more investment opportunities for enterprises in central cities.