Does High-Speed Railway Opening Improve the M&A Behavior?

Abstract

High-speed railway (HSR) shortens the spatial and temporal distance between regions and has a profound impact on regional economy and enterprise decision-making. This study investigates the impact of opening a high-speed railway (HSR) on enterprises’ mergers and acquisitions (M&A) behavior in China. Our findings suggest that the opening of HSR promotes the M&A activities of listed companies, and the opening of non-intercity HSR has a more obvious effect on the promotion of M&A decisions of enterprises. The results were robust after a series of robustness checks. Hence, the spatiotemporal squeezing effect generated by opening an HSR significantly improves the efficiency of information exchange and decreases the transaction costs of listed companies, which greatly promotes the M&A decisions of enterprises.

1. Introduction

M&A of listed companies is a major matter that affects the vital interests of various stakeholders and is crucial to promote industry integration. Regarding the influencing factors of M&A behavior, existing studies focus on external influences such as national industrial policies, enterprises’ political connection, and product market competition, as well as internal factors such as board networks and executive incentive. Taking geographical distance into account, there are significant discrepancies in information transmission speed and factor circulation efficiency across listed companies in different locations. The transaction costs associated with the geographical distance between enterprises and other stakeholders have been demonstrated to have a significant impact on their business performance (Kalnins and Lafontaine, 2013) [1], as well as their investment and financing behavior (Kang and Kim, 2008) [2]. The greater the geographical distance between the acquirer enterprise and the target enterprise, the higher the cost of information collection and the more difficult it is for the acquirer enterprise to comprehend the target company’s genuine operations (Ragozzino, 2009) [3], by which M&A behavior is affected. On the other hand, with the increase of the distance between listed enterprises and stakeholders such as investors, auditors, suppliers, analysts, and media, the cost of supervision becomes higher and the difficulty in restricting investment risks from agency problem becomes more aggravated (John et al., 2011) [4]. Therefore, the transaction costs associated with geographical distance influence enterprises’ M&A behavior.

Using the difference-in-difference (DID) model, this paper examines the impact of the quasi-natural experimental event of HSR opening on enterprises’ M&A behavior in China. The results show that the opening of HSR can promote the M&A activities of listed companies in the corresponding prefecture-level administrative districts. These results remain robust after a series of robustness tests, including parallel trend tests, endogeneity tests, replacing key variables, controlling the impact of transportation infrastructure construction, and excluding samples of large cities. Further examination shows that the opening of non-intercity HSR has a more obvious effect on the promotion of M&A decisions of enterprises.

There are three potential contributions of this paper. Firstly, using the quasi-natural experiment of HSR, this paper investigates the impact of improved transportation circumstances on companies’ M&A behavior. This is a new viewpoint on the drivers of M&A behavior. Secondly, this paper examines the impact of opening HSR on M&A behavior, which provides new firm-level empirical evidence based on the microeconomic consequences of HSR opening. Thirdly, taking the opening of HSR as a start point, our research expands the scope of using new economic geography in the micro-field of enterprises (e.g., corporate behavior).

2. Literature Review and Hypothesis Development

2.1. Literature Review

2.1.1. The Impact of Geographical Distance on Enterprise M&A Behavior

According to the new economic geography theory, geographical distance affects the decisions of economic agents by influencing their transaction costs, which also leads to differences in the spatial distribution of economic activities (Krugman, 1994) [5]. As a result, listed companies will also focus on transaction expenses due to geographical distance when carrying out M&A. Studies have shown that due to the requirement for information collection and ongoing communication, listed firms seek to invest in places with lower information collection costs and transportation expenses (Giroud, 2013) [6].

Based on the above logic, studies have been conducted to explore the impact of geographical distance on enterprise M&A behavior. Ragozzino (2009) found that there is less information asymmetry between geographically proximate companies [3]. So, the acquirer company is better informed about the operations of the target company. As a result, listed companies are more likely to acquire companies that are closer to their locations (Kang and Kim, 2008) [2]. On the other hand, closer physical proximity makes supervision of subsidiary management by the parent business easier after M&A, resulting in more effective integration (Chakrabarti and Mitchell, 2013) [7]. Di Guardo et al. (2016) found through further research that greater geographical distance brings higher information asymmetry, which increases cultural differences as well as political differences. This has a significant negative impact on M&A [8]. The above literature review shows that transaction costs associated with geographical factors are important factors that economic agents consider when making M&A decisions. However, the existing literature has only explored the economic consequences of geographical distance and has paid less attention to the impact of human geography, such as transportation, on enterprise M&A behavior.

2.1.2. Economic Consequences of the Opening of HSR

HSR is one of the most important infrastructure developments. Infrastructure developments profoundly affect business sustainability, for example, the Shanghai Hongqiao district effect proposed by Chopra et al. (2021) [9]. The transportation infrastructure has brought various economic agents closer together and modified the spatial structure between cities (Clark, 1998) [10]. HSR, as a product of transportation development, while solving the problem of rapid transportation of a large number of passengers, has affected the efficiency of transmission of information and other elements. As a result, the economic consequences of the opening of HSR have received wide attention from both academic and practical circles. The economic consequences of the opening of HSR have been studied in the literature from two aspects: regional economy and company decision-making.

Different studies have reached different conclusions regarding the impact of HSR opening on the regional economy. It has been proved that the opening of HSR lowers transportation costs, thus boosting regional economic growth and wage growth (Yin et al., 2015) [11], leading to an increase in regional income levels (Chen and Silva, 2013) [12]. Liang et al. (2020) concluded that the opening of HSR can promote the economic growth of underdeveloped regions, which is mainly achieved by pushing investment and optimizing the industrial structure, in addition to providing favorable conditions for the development of tertiary industries along the route [13]. Wang and Cai (2020) show that HSR opening in more developed cities boosts the innovation capability of the less developed cities around them greatly [14]. On the other hand, some studies suggest that the "siphon effect" created by the opening of HSR only benefits the development of central cities or large cities, which will have an impact on the interests of the surrounding areas and magnify regional disparities in development (Vickerman et al., 1999) [15]. Li et al. (2020) found that large cities get faster economic growth and higher resource allocation efficiency with the siphon effect due to the opening of HSR [16].

In terms of the impact of HSR on business decisions, studies have found that the opening of HSR reduces firm inventory levels by declining in transportation and communication cost, as well as agglomeration effect (Cui and Li, 2019) [17]. Furthermore, Yang et al. (2019) demonstrated that the opening of HSR can optimize the efficiency of resource allocation between core and surrounding cities, thereby positively affecting core city productivity while negatively affecting peripheral city productivity [18]. Duan et al. (2020) found that the opening of HSR improves transportation accessibility and thus significantly boosts venture capital investment mobility across cities [19].

According to the above literature review, existing studies generally agree that the opening of HSR can reduce the transaction costs associated with geographical distance and facilitate the production factor mobility, which has an impact on the regional economy and company decision-making. However, less research has focused on M&A as a critically important resource allocation strategy.

2.1.3. The Influencing Factors of Enterprise M&A Behavior

Enterprise M&A behaviors are important matters and important decisions that affect the interests of stakeholders. One of the main concerns in both theory and practice is what factors can have a substantial impact on enterprise M&A behavior. So far, studies have been conducted to investigate the factors that drive corporate M&A behavior, primarily in terms of external and internal factors.

About the external factors, studies have investigated the impact of national industrial policies (Barbieri et al., 2021) [20], enterprises’ political connection (Schweizer et al., 2019) [21] and product market competition (Lee et al., 2019) [22] on enterprise M&A behavior. Barbieri et al. (2021) suggested that enterprise M&A behavior may not just be related to strategic individual behaviors activated by firms, but also stimulated by governments as a tool to promote structural changes in the sectors’ market [20]. Schweizer et al. (2019) showed that executives with a political connection are more likely to complete cross-border M&A at the expense of shareholders [21]. Lee et al. (2019) found that in an emerging economy with less developed capital markets and insufficient investor protection, product market competition improves market efficiency for corporate control in an emerging economy, supporting the complementary hypothesis between product market competition and corporate takeover [22].

About the external factors, studies have investigated the impact of board networks (Renneboog and Zhao, 2014; Cai and Sevilir, 2012) [23,24] and executive incentive (Grinstein and Hribar, 2004; Zhao et al., 2016) [25,26] on enterprise M&A behavior. Renneboog and Zhao (2014) investigated the impact of social network on the corporate M&A process from the perspective of board networks, finding that friendly social contacts make the acquirer company more active in the M&A process and that the information gathering ability of directors and interlocking directors can improve the success rate of M&A while shortening the negotiation time [23]. In addition, Cai and Sevilir (2012) found that both transactions with a first-degree connection and a second-degree connection are able to generate higher revenues for the acquirer company, indicating that board connectedness plays important roles in M&A behavior and leads to greater value creation [24]. Based on an executive incentive perspective, Grinstein and Hribar (2004) showed that CEOs receive higher bonus incentives when M&A deals are larger, and this incentive makes CEOs put more effort into closing M&A agreements [25]. In addition, Zhao et al. (2016) explored the key factors that influence the performance of corporate M&A in state-owned enterprises and found that increasing executive compensation can significantly improve the post-merger profitability of state-owned enterprises [26].

The above literature review shows that studies have explored the factors influencing enterprise M&A behavior in terms of external factors such as national industrial policies, enterprises’ political connection and product market competition, as well as internal factors such as board networks and executive incentive. However, studies have ignored the impact of objective environment and its changes on the M&A behavior of listed companies. Due to China’s vast territory, geographical distance, and other objective circumstances, there are significant discrepancies in information transmission speed and factor circulation efficiency across listed companies in different locations. Therefore, this paper delves into the impact of the objective environmental changes triggered by the opening of HSR on enterprise M&A behavior.

2.2. Hypothesis Development

M&A transactions are an important way for companies to integrate and upgrade their industries, and the success of M&A transactions will have a profound impact on the value of the company (King et al., 2004) [27]. New geography economics argues that the transaction costs induced by geographical distance significantly influence the decision-making behavior of economic agents, thus shaping the spatial distribution of economic activities (Krugman, 1994) [5]. The cost of information gathering for investors rises dramatically as geographical distance increases, and their capacity to acquire private information decreases, affecting their M&A behavior. The opening of HSR has reorganized city systems and regional economies, and can accelerate the flow of information and other factors between cities, lowering transaction costs and facilitating corporate M&A.

Firstly, the opening of HSR decreases the time and expense costs of travel, allowing the acquirer company to find more M&A targets, thus enhancing the likelihood of M&A activities. Directors with investment banks working experience have been found to reduce transaction costs for MNCs, find more and more suitable M&A targets for MNCs, and hence conduct better M&A transactions (Huang et al., 2014) [28]. With the opening of HSR, the acquirer company can now conduct on-site inspections in a shorter time and at a lower cost before the M&A, allowing for the identification of more M&A targets on a wider scale. This will significantly increase the probability of finding the right M&A target and increase the likelihood of an M&A event occurring.

Secondly, the opening of HSR reduces the cost of information gathering and communication, which helps the acquirer company to better balance the interests of all parties when formulating an M&A strategy, boosting the likelihood of M&A. It has been shown that an effective information communication system can alleviate the degree of information asymmetry between the acquirer and the target and lead to better decisions (Raman et al., 2013) [29]. Renneboog and Zhao (2014) investigated the impact of social networks on enterprise M&A process, friendly social contacts make the acquirer company more active in the M&A process, and the information gathering ability of directors and interlocking directors can improve the success rate of M&A while shortening the negotiation time [23]. The opening of HSR allows the acquirer company to conduct on-site investigations of the M&A target more frequently at a lower cost, allowing it to obtain soft information such as asset quality, profitability, and development prospects of the target company while also verifying hard information such as financial reports. As a result, the degree of information asymmetry between the acquirer and the target is effectively reduced. When the acquirer company has more comprehensive information, it aids organizations in reducing the uncertainty problems they face during the M&A process, making M&A activities more smoothly (Xia et al., 2014) [30]. The following hypothesis is thus proposed:

Hypotheses 1 (H1).

The opening of HSR promotes enterprise M&A activities.

In detail, the National Railway Administration of the People’s Republic of China separates HSR into intercity and non-intercity HSR on their website. Intercity HSR refers to HSR built in densely populated metropolitan areas and urban planning zones characterized by short operating distances and public transport characteristics. In this paper, non-intercity HSR is an HSR that excludes intercity HSR and operates at relatively long distances.

The previous analysis in this paper shows that, on the one hand, the opening of HSR decreases the time and expense of travel, allowing the acquirer company to find more M&A targets, thus enhancing the likelihood of M&A activities; on the other hand, the opening of HSR reduces the cost of information gathering and communication, which helps the acquirer company to better balance the interests of all parties when formulating an M&A strategy, boosting the likelihood of M&A.

First, due to the short operating distance of intercity HSR, its ability to expand the scope of M&A of the acquirer company is limited, and it is difficult to help the acquirer company to find more M&A targets. The longer operating mileage of non-intercity HSR is more likely to help the acquirer company to expand the scope of M&A and find M&A targets, which can effectively improve the probability of making M&A decisions. Second, intercity HSR has a small coverage area and is usually only able to influence local stakeholders. Since local stakeholders already have an information advantage, the impact of intercity HSR opening on information gathering is limited. The opening of non-intercity HSR can reduce the search and communication costs of distant stakeholders and alleviate their information disadvantage. In summary, we conclude that the opening of non-intercity HSR has a more significant effect on the probability of enterprises carrying out M&A decisions than intercity HSR. Moreover, the effect should be more pronounced if both kinds of HSR are opened at the same time. Thus, the following hypothesis is proposed:

Hypotheses 2 (H2).

Compared to intercity HSR, the opening of non-intercity HSR has a more positive impact on enterprise M&A activities.

3. Methodology

3.1. Sample Selection and Data Sources

Using the DID model (Viglia et al., 2021) [31] and a sample of enterprises listed in Shanghai Stock Exchange and Shenzhen Stock Exchange from 2006 to 2019, we examine the impact of the quasi-natural experimental event of HSR opening on enterprises’ M&A behavior (Lim, 2021) [32]. The treatment group is a sample of companies that opened the HSR at the location level between 2006 and 2019. The control group is a sample of companies without HSR at the location between 2006 and 2019, and a sample of companies before the opening of the HSR. The final sample has 26,899 firm-year observations, involving 3496 listed companies in 270 prefectural-level cities in China.

The basic information on M&A events used in this paper was obtained from the WIND database of M&A and restructuring of listed companies and the CSMAR database of M&A. The data on the opening of HSR were manually compiled from the website of China National Railway Group Limited. Specifically, we first query the data of each city HSR station (city name—station name—HSR line name) (The number of HSR stations and the number of HSR lines passed by each city in each year in the robustness tests are derived from this.) from the website of China Railway Corporation and the website of China National Railway Group Limited, and then determined the time when each station first opened HSR through search engines such as Baidu, and finally obtained the data of the year when each city first opened HSR. GDP per capita of prefecture-level administrative districts are manually sorted from the China City Statistical Yearbook. Local total value of imports and exports and local unemployment rate are manually sorted from the China Statistical Yearbook. Scores of the marketization index are manually sorted from the NERI Index of Marketization of China’s Provinces. The companies’ financial data were obtained from the WIND and CSMAR databases. To avoid being influenced by outliers, we winsorize all the continuous variables at the 1% and 99% levels.

3.2. Variable Definition

3.2.1. Dependent Variable

M&A is a dummy variable with the value of 1 if there is an M&A in the current year and 0 otherwise.

3.2.2. Independent Variable

If the prefecture-level cities where the company locates has not yet opened a HSR during the sample period, HSR is defined to the value of 0.

If the prefecture-level cities where the company locates has opened a HSR during the sample period, the value of HSR is 0 before opening a HSR and 1 after opening HSR.

3.3. Model Construction

We develop model (1) to examine the impact of opening of HSR on enterprises’ M&A decisions in order to test H1.

$$M\&A_{i,t}={\beta }_0+{\beta }_{1}HS{R}_{i,t}+{\beta }_{2}Contro{l}_{i,t}+{\epsilon }_{i,t}$$

(1)

The dependent variable M&A and the independent variable HSR are explained above. According to Borthwick et al. (2020) [33], firm controls include company size (Size), return on assets (ROA), debt to assets ratio (Leverage), growth capacity (Growth), operating cash flow (OCF), Tobin Q (Tobin’s Q), board size (Board), percentage of independent directors (Independ), equity concentration (Equity), dual ownership (Dual), managerial stockholding level (ESH), state ownership (SOE), market to book ratio (M/B), annualized monthly returns (PastReturns), standard deviation of monthly returns (Volatility), economic development (LnGDP), economic opening (Open), and unemployment rate (Unemploy). The definitions of all variables are presented in

Table 1. Definition of variables.

Variables	Definition
M&A	An indicator variable that equals 1 if the company made an M&A during the year, otherwise it equals 0.
HSR	If the prefecture-level administrative district where the company is located has not yet opened a HSR during the sample period, HSR equals 0. If the prefecture-level administrative district where the company is located has opened a HSR during the sample period, HSR equals 0 before opening a HSR and 1 after opening a HSR.
ChenJi	When the intercity HSR is opened in the prefecture-level administrative district where the company is located, ChenJi equals 1, otherwise it is 0.
NonChenJi	When the non-intercity HSR is opened in the prefecture-level administrative district where the company is located, NonChenJi equals 1, otherwise it is 0.
TwoTrain	When the intercity HSR and non-intercity HSR are opened in the prefecture-level administrative district where the company is located at the same time, TwoTrain equals 1, otherwise it is 0.
Size	Logarithmic of the company’s assets one year before M&A.
ROA	Company’s return on assets one year before M&A.
Leverage	Company’s debt to assets ratio one year before M&A.
Growth	Growth rate of operating income one year before M&A.
OCF	Operating cash flow / total assets, data from one year before M&A.
Tobin’ s Q	(Equity market value + net debt market value) / total assets, data from one year before M&A.
Board	Total number of board members one year before M&A.
Independ	Number of independent directors / total number of board members, data from one year before M&A.
Equity	Percentage of shareholding of the largest shareholder one year before M&A.
Dual	An indicator variable that equals 1 if the CEO is also the chairman of the board, otherwise it equals 0.
ESH	Number of shares held by management / total number of shares, data from one year before M&A.
SOE	Company’s state ownership one year before M&A.
M/B	Market value of equity divided by book value of equity.
PastReturns	Annualized monthly returns for the calendar year.
Volatility	The standard deviation of firm-level monthly returns.
LnGDP	GDP per capita of prefecture-level administrative district where company is located, data from one year before M&A.
Open	Ratio of total import and export of goods to GDP by domestic destination and origin in the province where the company is located, data from one year before M&A.
Unemploy	Unemployment rate of prefecture-level administrative district where company is located, data from one year before M&A.
Year	Year indicators.
Industry	Industry indicators.

. In addition, we account for year and industry fixed effects in model (1).

To test H2, we distinguish HSR opening (HSR) into intercity HSR opening (ChenJi), non-intercity HSR opening (NonChenJi) and both types of HSR opening (TwoTrain). We use ChenJi, NonChenJi, and TwoTrain to replace HSR in model (1) for regression. The definitions are similar to HSR. When the intercity HSR (or non-intercity HSR, or both intercity and non-intercity HSR) is opened in the prefecture-level administrative district where the company is located, ChenJi (or NonChenJi, or TwoTrain) takes the value of 1, otherwise it is 0.

4. Main Results and Discussions

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variables	N	Mean	Std Dev.	Minimum	Median	Maximum	Range
M&A	26,899	0.3212	0.4669	0.0000	0.0000	0.0000	0.0000
HSR	26,899	0.7347	0.4415	0.0000	0.0000	1.0000	0.0000
ChenJi	26,899	0.2227	0.4161	0.0000	0.0000	1.0000	1.0000
NonChenJi	26,899	0.6949	0.4605	0.0000	1.0000	1.0000	1.0000
TwoTrain	26,899	0.1825	0.3863	0.0000	0.0000	1.0000	1.0000
Size	26,899	21.9348	1.2609	19.3756	21.0270	21.7782	2.4026
ROA	26,899	0.0365	0.0605	−0.2525	0.0135	0.0360	0.2885
Leverage	26,899	0.4456	0.2134	0.0510	0.2763	0.4438	0.3928
Growth	26,899	0.4485	1.3584	−0.7398	−0.0389	0.1318	0.8716
OCF	26,899	0.0434	0.0743	−0.1912	0.0034	0.0431	0.2343
Tobin’ s Q	26,899	2.0238	1.3219	0.9043	1.2220	1.5870	0.6827
Board	26,899	2.2665	0.1800	1.7918	2.1972	2.3026	0.5108
Independ	26,899	0.3701	0.0523	0.2857	0.3333	0.3333	0.0476
Equity	26,899	0.3518	0.1499	0.0877	0.2329	0.3308	0.2431
Dual	26,899	0.2361	0.4247	0.0000	0.0000	0.0000	0.0000
ESH	26,899	0.1116	0.1907	0.0000	0.0000	0.0004	0.0004
SOE	26,899	0.4272	0.4947	0.0000	0.0000	0.0000	0.0000
M/B	26,899	2.0590	1.3523	0.8819	1.6131	8.8898	8.0079
PastReturns	26,899	0.0024	0.0073	−0.0071	0.0004	0.0318	0.0389
Volatility	26,899	0.0013	0.0006	0.0004	0.0012	0.0035	0.0031
LnGDP	26,899	11.2147	0.6137	9.4331	10.8611	11.2942	1.8611
Open	26,899	0.5034	0.4180	0.0000	0.1396	0.4126	0.4126
Unemploy	26,899	0.0920	0.1251	0.0130	0.0300	0.0350	0.0220

reports the descriptive statistics results for the main variables. As shown in the table, the mean of corporate M&A decisions (M&A) is 0.3212, implying that 32 out of 100 companies have completed M&A. The mean value of HSR opening (HSR) is 0.7347, suggesting that HSR is available in 73.47 percent of the sample’s prefecture-level administrative districts at the office location level.

4.2. Empirical Results

We investigate the impact of HSR opening on M&A decisions in this part. Columns (1) and (2) show the results of regressions using the Probit model, and columns (3) and (4) show the results of regressions using the fixed effects model, columns (1) and (3) contain only explanatory and explained variables, while columns (2) and (4) further add control variables. The results in

Table 3. Impact of the opening of HSR on M&A behavior.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
HSR	0.1828 ***	0.1379 ***	0.0510 ***	0.0411 ***
	(5.77)	(4.23)	(4.94)	(3.93)
Size		0.2573 ***		0.0825 ***
		(18.30)		(17.66)
ROA		1.1529 ***		0.3188 ***
		(5.99)		(5.97)
Leverage		−0.1165 *		−0.0623 ***
		(−1.84)		(−3.13)
Growth		0.0084		0.0028
		(1.20)		(1.18)
OCF		−0.3732 ***		−0.0943 **
		(−2.81)		(−2.25)
Tobin’ s Q		0.0691 ***		0.0246 ***
		(5.86)		(6.50)
Board		0.0079		−0.0015
		(0.11)		(−0.06)
Independ		0.1862		0.0217
		(0.83)		(0.30)
Equity		−0.0246		0.0235
		(−0.32)		(0.89)
Dual		0.0316		0.0085
		(1.29)		(1.03)
ESH		0.3833 ***		0.1290 ***
		(5.65)		(5.56)
SOE		−0.1825 ***		−0.0548 ***
		(−6.58)		(−5.95)
M/B		−0.0752 ***		−0.0185 ***
		(−6.10)		(−4.91)
PastReturns		20.1783 ***		6.3737 ***
		(9.63)		(9.17)
Volatility		144.0080 ***		42.5956 ***
		(7.13)		(6.26)
LnGDP		0.0274		0.0115
		(0.99)		(1.28)
Open		−0.0281		−0.0104
		(−0.95)		(−1.12)
Unemploy		−0.5034 ***		−0.1415 ***
		(−3.26)		(−2.84)
Constant	−0.9449 ***	−6.8573 ***	0.2000 ***	−1.7080 ***
	(−10.46)	(−15.61)	(6.50)	(−11.95)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.0189	0.0596	0.0231	0.0719

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that opening of HSR (HSR) is significantly and positively related to corporate M&A decisions (M&A) at the 1% significance level. The above findings indicate that HSR opening aids in company M&A decision-making, confirming research hypothesis H1.

In terms of control variables, company size (Size), return on assets (ROA), managerial stockholding level (ESH), annualized monthly returns (PastReturns), and standard deviation of monthly returns (Volatility) are significantly and positively related to M&A decisions. This demonstrates that the larger the company, the higher the return on assets, the higher the shareholding ratio of management, the higher the annualized monthly returns, the higher the standard deviation of monthly returns, the more inclined the listed companies are to make M&A decisions. Operating cash flow (OCF), state ownership (SOE), market to book ratio (M/B) and unemployment rate (Unemploy) are significantly and negatively related to M&A decisions. This demonstrates that the higher the cash flow, the more likely the enterprise is state-owned, the higher the market to book ratio, the higher the unemployment rate, and the less likely the listed companies tend to make M&A decisions.

The results presented in

Table 4. Impact of different HSR types on M&A behavior.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
ChenJi	−0.0059	−0.0265	−0.0068	−0.0147
	(−0.11)	(−0.50)	(−0.38)	(−0.85)
NonChenJi	0.1318 ***	0.1141 ***	0.0378 ***	0.0329 ***
	(4.19)	(3.54)	(3.65)	(3.16)
TwoTrain	0.1726 ***	0.1340 **	0.0643 ***	0.0550 ***
	(3.02)	(2.39)	(3.33)	(2.93)
Size		0.2558 ***		0.0819 ***
		(18.19)		(17.56)
ROA		1.1388 ***		0.3130 ***
		(5.90)		(5.85)
Leverage		−0.1122 *		−0.0615 ***
		(−1.78)		(−3.10)
Growth		0.0080		0.0027
		(1.14)		(1.14)
OCF		−0.3509 ***		−0.0876 **
		(−2.64)		(−2.09)
Tobin’ s Q		0.0684 ***		0.0244 ***
		(5.80)		(6.45)
Board		0.0060		−0.0022
		(0.08)		(−0.09)
Independ		0.1631		0.0156
		(0.73)		(0.21)
Equity		−0.0215		0.0249
		(−0.28)		(0.95)
Dual		0.0326		0.0085
		(1.33)		(1.04)
ESH		0.3765 ***		0.1268 ***
		(5.55)		(5.47)
SOE		−0.1889 ***		−0.0572 ***
		(−6.80)		(−6.21)
M/B		−0.0751 ***		−0.0184 ***
		(−6.09)		(−4.89)
PastReturns		20.0039 ***		6.3092 ***
		(9.54)		(9.07)
Volatility		144.6108 ***		42.6667 ***
		(7.16)		(6.27)
LnGDP		0.0160		0.0075
		(0.57)		(0.82)
Open		−0.0105		−0.0049
		(−0.35)		(−0.52)
Unemploy		−0.3803 **		−0.1080 **
		(−2.45)		(−2.16)
Constant	−0.9295 ***	−6.6916 ***	0.2066 ***	−1.6491 ***
	(−10.24)	(−15.16)	(6.62)	(−11.50)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.0206	0.0604	0.0254	0.0730

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that intercity HSR opening (ChenJi) is not significantly related to the enterprise M&A decisions (M&A), while non-intercity HSR opening (NonChenJi) and both types of HSR opening (TwoTrain) are significantly and positively related to enterprises’ M&A decisions (M&A) at the 1% significance level. The results indicate that the opening of non-intercity HSR has a more significant effect on the probability of enterprises making M&A decisions than intercity HSR.

5. Robustness Tests

5.1. Parallel Trend Test

To test whether the sample in this paper satisfies the parallel trend hypothesis, we refer to the study of Bertrand and Mullainathan (2003) [34] and define five year dummy variables according to the following method. Before2⁺, Before1 represent 1 and 2 years before the opening of HSR, Present represents the year when HSR opened, After1, After2⁺ represent 1 and 2 years after the opening of HSR. The coefficients of model (2) are then calculated by multiplying the dummy variables for each year by the dummy variable for whether or not the HSR was open (HSRDum).

$$\begin{gathered} M\&A_{i,t}\left(BHA{R}_{i.t}\right)={\beta }_0+{\beta }_{1}Before{2}^{+}\ast HSRDu{m}_{i,t}+{\beta }_{2}Before1\ast HSRDu{m}_{i,t} \\ +{\beta }_{3}Present\ast HSRDu{m}_{i,t}+{\beta }_{4}After1\ast HSRDu{m}_{i,t}+{\beta }_{5}After{2}^{+}\ast HSRDu{m}_{i,t} \\ +{{\displaystyle \sum }}^{ }Contro{l}_{i,t}+{{\displaystyle \sum }}^{ }Year+{{\displaystyle \sum }}^{ }Industry+{\epsilon }_t \end{gathered}$$

(2)

The results in

Table 5. Parallel trend test.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
Before2+*HSRDum	0.0914	0.0489	0.0369	0.0189
	(1.36)	(0.77)	(1.58)	(0.90)
Before1*HSRDum	0.0816	0.0260	0.0297	0.0097
	(1.16)	(0.38)	(1.22)	(0.43)
Present*HSRDum	0.2034 ***	0.1534 **	0.0716 ***	0.0509 **
	(2.97)	(2.29)	(2.98)	(2.27)
After1*HSRDum	0.2280 ***	0.1644 **	0.0772 ***	0.0532 **
	(3.35)	(2.47)	(3.24)	(2.40)
After2+*HSRDum	0.2772 ***	0.1805 ***	0.0877 ***	0.0571 ***
	(4.45)	(2.95)	(4.07)	(2.82)
Size		0.2572 ***		0.0824 ***
		(18.30)		(17.66)
ROA		1.1513 ***		0.3184 ***
		(5.99)		(5.97)
Leverage		−0.1167 *		−0.0622 ***
		(−1.84)		(−3.13)
Growth		0.0084		0.0028
		(1.19)		(1.18)
OCF		−0.3729 ***		−0.0942 **
		(−2.81)		(−2.25)
Tobin’ s Q		0.0691 ***		0.0246 ***
		(5.86)		(6.50)
Board		0.0086		−0.0013
		(0.12)		(−0.06)
Independ		0.1885		0.0224
		(0.84)		(0.31)
Equity		−0.0247		0.0235
		(−0.32)		(0.89)
Dual		0.0312		0.0084
		(1.27)		(1.02)
ESH		0.3822 ***		0.1286 ***
		(5.62)		(5.54)
SOE		−0.1822 ***		−0.0547 ***
		(−6.56)		(−5.94)
M/B		−0.0751 ***		−0.0185 ***
		(−6.09)		(−4.91)
PastReturns		20.1559 ***		6.3656 ***
		(9.61)		(9.15)
Volatility		143.6729 ***		42.5231 ***
		(7.11)		(6.25)
LnGDP		0.0259		0.0110
		(0.91)		(1.19)
Open		−0.0289		−0.0109
		(−0.97)		(−1.17)
Unemploy		−0.5047 ***		−0.1421 ***
		(−3.26)		(−2.85)
Constant	−1.0248 ***	−6.8839 ***	0.1675 ***	−1.7192 ***
	(−9.38)	(−15.40)	(4.55)	(−11.84)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.0191	0.0597	0.0234	0.0719

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that the coefficients of Before²⁺*HSRDum and Before1*HSRDum are not significant. The trends of the treatment and control groups are not significantly different before the opening of HSR. Present*HSRDum, After1*HSRDum, and After2⁺*HSRDum are all significantly and positively related to enterprise M&A decisions (M&A), indicating that the effect of HSR opening on corporate M&A decisions is significant both in the year of opening and after opening. These results suggest that the effect of HSR opening on M&A decisions is persistent.

5.2. Endogeneity Test

In order to further control the endogeneity problem, we select the topographic relief (Slope) of the prefecture-level administrative district where the listed company is located as an instrumental variable and use the Heckman two-stage model to solve the above problem.

The basis for selecting Slope as the instrumental variable is as follows. On the one hand, Slope fully reflects the topographic condition of a specific area and can directly affect the cost of HSR construction. In the process of HSR construction, the cost of building HSR in areas with low topographic relief is much lower than that in areas with high topographic relief, and the current HSR lines in China are mainly concentrated in areas with low topographic relief such as plains. Therefore, topographic relief is one of the important factors affecting the construction of HSR, and this indicator satisfies the requirement that the instrumental variables should be correlated. On the other hand, topographic relief, as an important representative of terrain conditions, is a geographical condition that has evolved over millions of years and is not directly related to the enterprises’ M&A behavior. This indicator satisfies the requirement that the instrumental variable should be exogenous.

Referring to the study of Zhang et al. (2019) [35], the topographic relief (Slope) of the prefecture-level administrative district where the listed company is located is extracted mainly by the window method using the GRID and TABLE modules of ARC/INFO. Practically, we set 10 km by 10 km windows in the extent of the Chinese mainland and calculate the topographic relief according to the formula below.

$$Slope=\left\{\left[Max\left(H\right)-Min\left(H\right)\right]\times \left[1-P\left(A\right)/A\right]\right\}/500$$

(3)

Slope is the topographic relief, the ratio of regional elevation difference to national elevation difference; Max(H) and Min(H) are the highest and lowest elevations in the region, P(A) is the average area of the region; A is the total area of the region. We adopt a window of 10km by 10km as the cell; the value of A is 10 × 10 = 100 km². 500m is the base elevation of low mountains in China’s landform type.

In the first stage of the Heckman two-stage model, we conduct a Probit regression with the opening of HSR (HSR) as the dependent variable along with the control variables in the previous section. Moreover, the average topographic relief (Slope) of the prefecture-level administrative district where the listed company is located is used as an instrumental variable in this section along with the regression.

In the second stage of the Heckman two-stage model, the inverse Mills ratio (λ) from the first stage above is added to the regression. The results are shown in

Table 6. Endogeneity test (Heckman two-stage model).

Variables	Probit	Heckman
	HSR	M&A
		Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)	(5)
HSR		0.1512 ***	0.1463 ***	0.0487 ***	0.0441 ***
		(4.52)	(4.32)	(4.39)	(4.04)
Size	0.0380		0.2561 ***		0.0818 ***
	(1.36)		(17.34)		(16.63)
ROA	−0.2385		1.1470 ***		0.3304 ***
	(−0.68)		(5.69)		(5.84)
Leverage	0.0755		−0.1277 *		−0.0651 ***
	(0.56)		(−1.93)		(−3.10)
Growth	0.0207 *		0.0089		0.0027
	(1.74)		(1.21)		(1.12)
OCF	0.0909		−0.3019 **		−0.0669
	(0.41)		(−2.18)		(−1.51)
Tobin’ s Q	0.0198		0.0716 ***		0.0262 ***
	(1.17)		(5.85)		(6.58)
Board	−0.1014		0.0354		0.0096
	(−0.65)		(0.46)		(0.38)
Independ	0.3882		0.2133		0.0136
	(0.85)		(0.92)		(0.18)
Equity	0.1838		−0.0136		0.0261
	(1.14)		(−0.17)		(0.95)
Dual	−0.0494		0.0242		0.0062
	(−0.96)		(0.96)		(0.74)
ESH	0.4700 ***		0.3851 ***		0.1286 ***
	(3.10)		(5.58)		(5.43)
SOE	0.1378 **		−0.1855 ***		−0.0572 ***
	(2.19)		(−6.39)		(−5.87)
M/B	0.0198		−0.0782 ***		−0.0211 ***
	(1.10)		(−5.99)		(−5.27)
PastReturns	3.8589		24.0893 ***		8.2070 ***
	(1.15)		(9.84)		(9.86)
Volatility	−11.7969		157.1176 ***		46.6506 ***
	(−0.33)		(7.48)		(6.51)
LnGDP	1.1854 ***		0.0594 *		0.0244 **
	(16.39)		(1.65)		(2.08)
Open	−0.0772		−0.0106		−0.0042
	(−1.21)		(−0.33)		(−0.41)
Unemploy	−0.2155		−0.4988 ***		−0.1353 ***
	(−0.51)		(−3.20)		(−2.70)
Slope	−0.4070 ***
	(−10.72)
IMR		−0.0781 **	0.0756	−0.0103	0.0300 *
		(−2.03)	(1.61)	(−0.83)	(1.93)
Constant	−12.3688 ***	−0.3968 ***	−6.9726 ***	0.3335 ***	−1.7940 ***
	(−11.31)	(−3.30)	(−11.93)	(7.71)	(−9.28)
Industry Effect	YES	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.4468	0.0164	0.0572	0.0204	0.0700

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

. The results show that the instrumental variable (Slope) in the first-stage regression is significantly negative at the 1% significance level, indicating that the higher the regional slope, the lower the likelihood of HSR opening in the region. After adding the inverse Mills ratio (λ) from the first stage into the second stage for regression, we find that opening of HSR (HSR) is significantly and positively related to the company’s M&A decision (M&A) at the 1% significance level, which is consistent with the results of the previous test.

5.3. Alternative Key Variables

In order to ensure the sensitivity of the metric, we use the number of HSR stations and the number of HSR lines passed by the prefecture-level administrative district where the listed company is located in each year to re-measure the opening of HSR (HSR).

5.3.1. Use Number of HSR Stations as a Metric

The number of HSR stations (HSR_ZD) in the prefecture-level administrative district where the listed company is located directly determines the intensity of the impact of HSR opening on listed companies. Specifically, the more HSR stations in the prefecture-level administrative district where the listed company is located, the denser the HSR network is, so the greater the influence of the HSR network on the listed companies in that district. Therefore, we use HSR_ZD to re-examine model (1). The results presented in

Table 7. Using the number of HSR stations as alternative dependent variables.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
HSR_ZD	0.0731 ***	0.0389 **	0.0285 ***	0.0156 **
	(3.87)	(2.12)	(4.38)	(2.47)
Size		0.2569 ***		0.0823 ***
		(18.20)		(17.59)
ROA		1.1380 ***		0.3164 ***
		(5.91)		(5.93)
Leverage		−0.1155 *		−0.0619 ***
		(−1.82)		(−3.11)
Growth		0.0085		0.0028
		(1.21)		(1.19)
OCF		−0.3621 ***		−0.0913 **
		(−2.73)		(−2.18)
Tobin’ s Q		0.0689 ***		0.0246 ***
		(5.84)		(6.49)
Board		0.0039		−0.0024
		(0.05)		(−0.10)
Independ		0.1845		0.0229
		(0.82)		(0.32)
Equity		−0.0193		0.0248
		(−0.25)		(0.94)
Dual		0.0307		0.0080
		(1.25)		(0.98)
ESH		0.3876 ***		0.1301 ***
		(5.70)		(5.59)
SOE		−0.1811 ***		−0.0544 ***
		(−6.51)		(−5.91)
M/B		−0.0754 ***		−0.0186 ***
		(−6.11)		(−4.92)
PastReturns		20.3068 ***		6.4062 ***
		(9.68)		(9.21)
Volatility		144.3255 ***		42.6532 ***
		(7.15)		(6.27)
LnGDP		0.0397		0.0125
		(1.41)		(1.35)
Open		−0.0207		−0.0072
		(−0.70)		(−0.77)
Unemploy		−0.4845 ***		−0.1361 ***
		(−3.14)		(−2.73)
Constant	−1.0927 ***	−7.0543 ***	0.1476 ***	−1.7474 ***
	(−11.24)	(−16.06)	(4.50)	(−12.25)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.0183	0.0591	0.0224	0.0713

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that HSR_ZD is significantly and positively related to corporate M&A decisions (M&A) at the 1% or 5% significance level, which is consistent with the results of the previous tests.

5.3.2. Use Number of HSR Lines as a Metric

The number of HSR lines (HSR_XL) passing through the prefecture-level administrative district where the listed company is located also determines the intensity of the impact of the opening of HSR on listed companies in the region. Specifically, if several HSR lines have stations in the same prefecture-level administrative district, this district will form a node shaped like the letter X or Union Jack in the HSR network, signifying that the administrative district is located in a key area of the HSR network. Listed companies in such prefecture-level administrative districts will be more affected by the HSR network. Therefore, we use HSR_XL to re-examine model (1). The results presented in

Table 8. Using the number of HSR lines as alternative dependent variables.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
HSR_LX	0.1401 ***	0.0906 ***	0.0496 ***	0.0311 ***
	(5.65)	(3.60)	(5.86)	(3.59)
Size		0.2567 ***		0.0822 ***
		(18.22)		(17.58)
ROA		1.1278 ***		0.3147 ***
		(5.85)		(5.89)
Leverage		−0.1223 *		−0.0634 ***
		(−1.93)		(−3.19)
Growth		0.0082		0.0027
		(1.17)		(1.16)
OCF		−0.3527 ***		−0.0898 **
		(−2.66)		(−2.15)
Tobin’ s Q		0.0684 ***		0.0244 ***
		(5.79)		(6.45)
Board		0.0043		−0.0023
		(0.06)		(−0.10)
Independ		0.1746		0.0203
		(0.78)		(0.28)
Equity		−0.0265		0.0231
		(−0.34)		(0.88)
Dual		0.0327		0.0086
		(1.33)		(1.05)
ESH		0.3792 ***		0.1279 ***
		(5.58)		(5.50)
SOE		−0.1862 ***		−0.0561 ***
		(−6.70)		(−6.09)
M/B		−0.0755 ***		−0.0186 ***
		(−6.12)		(−4.93)
PastReturns		20.2666 ***		6.3934 ***
		(9.66)		(9.19)
Volatility		144.7902 ***		42.7169 ***
		(7.17)		(6.28)
LnGDP		0.0138		0.0052
		(0.47)		(0.53)
Open		−0.0118		−0.0053
		(−0.40)		(−0.56)
Unemploy		−0.4699 ***		−0.1331 ***
		(−3.04)		(−2.67)
Constant	−1.1196 ***	−6.8124 ***	0.1422 ***	−1.6758 ***
	(−11.83)	(−15.25)	(4.46)	(−11.49)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899
R2/Pseudo R2	0.0193	0.0595	0.0236	0.0718

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that HSR_XL is significantly and positively related to corporate M&A decisions (M&A) at the 1% significance level, which is consistent with the results of the previous tests.

5.4. Exclusion of Samples of Large Cities

To avoid political and economic factors of important cities from triggering endogenous problems related to the location of HSR, we exclude samples of provincial capitals, municipalities directly under the Central Government and municipalities with independent planning status, and conduct regression analysis. The results presented in

Table 9. Robustness test (H1): excluding samples of large cities.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
HSR	0.1608 ***	0.1369 ***	0.0452 ***	0.0391 ***
	(3.73)	(3.10)	(3.21)	(2.74)
Size		0.1742 ***		0.0549 ***
		(7.48)		(7.21)
ROA		1.4273 ***		0.4066 ***
		(4.39)		(4.44)
Leverage		−0.0565		−0.0412
		(−0.55)		(−1.29)
Growth		0.0262		0.0086
		(1.31)		(1.29)
OCF		−0.4942 **		−0.1221
		(−2.07)		(−1.63)
Tobin’ s Q		0.0383 *		0.0150 **
		(1.75)		(2.26)
Board		−0.0672		−0.0252
		(−0.54)		(−0.67)
Independ		−0.7569 **		−0.2613 **
		(−2.01)		(−2.31)
Equity		−0.0267		−0.0136
		(−0.21)		(−0.33)
Dual		0.0381		0.0007
		(0.98)		(0.05)
ESH		0.2246 **		0.0797 **
		(2.10)		(2.23)
SOE		−0.1485 ***		−0.0468 ***
		(−3.23)		(−3.15)
M/B		−0.0884 ***		−0.0212 ***
		(−3.94)		(−3.16)
PastReturns		21.6261 ***		6.5700 ***
		(6.05)		(5.68)
Volatility		158.6561 ***		47.5342 ***
		(4.68)		(4.24)
LnGDP		−0.0057		−0.0007
		(−0.15)		(−0.06)
Open		0.0308		0.0137
		(0.55)		(0.77)
Unemploy		−0.1573		−0.0239
		(−0.41)		(−0.21)
Constant	−1.0219 ***	−4.3953 ***	0.1668 ***	−0.8820 ***
	(−7.73)	(−6.39)	(4.36)	(−4.06)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	9818	9818	9818	9818
R2/Pseudo R2	0.0178	0.0426	0.0215	0.0494

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that HSR is still significantly and positively related to corporate M&A decisions (M&A) at the 1% significance level, which is consistent with the previous results.

The results presented in

Table 10. Robustness test (H2): excluding samples of large cities.

Variables	M&A
	Probit Model		Fixed Effects Model
	(1)	(2)	(3)	(4)
ChenJi	0.0035	−0.0265	−0.0009	−0.0106
	(0.04)	(−0.34)	(−0.03)	(−0.41)
NonChenJi	0.1296 ***	0.1089 **	0.0334 **	0.0282 *
	(2.98)	(2.46)	(2.32)	(1.95)
TwoTrain	0.2462 **	0.2377 **	0.0960 ***	0.0965 ***
	(2.42)	(2.40)	(2.72)	(2.80)
Size		0.1737 ***		0.0548 ***
		(7.45)		(7.20)
ROA		1.3936 ***		0.3966 ***
		(4.27)		(4.31)
Leverage		−0.0690		−0.0458
		(−0.67)		(−1.43)
Growth		0.0272		0.0088
		(1.36)		(1.34)
OCF		−0.4977 **		−0.1233 *
		(−2.09)		(−1.65)
Tobin’ s Q		0.0401 *		0.0155 **
		(1.83)		(2.33)
Board		−0.0633		−0.0247
		(−0.51)		(−0.66)
Independ		−0.7712 **		−0.2644 **
		(−2.06)		(−2.35)
Equity		−0.0183		−0.0106
		(−0.15)		(−0.25)
Dual		0.0348		−0.0007
		(0.89)		(−0.06)
ESH		0.2324 **		0.0823 **
		(2.17)		(2.30)
SOE		−0.1505 ***		−0.0475 ***
		(−3.28)		(−3.21)
M/B		−0.0880 ***		−0.0209 ***
		(−3.92)		(−3.11)
PastReturns		21.5412 ***		6.5358 ***
		(6.01)		(5.64)
Volatility		159.2026 ***		47.4819 ***
		(4.70)		(4.24)
LnGDP		0.0060		0.0028
		(0.16)		(0.23)
Open		0.0361		0.0154
		(0.64)		(0.86)
Unemploy		−0.0501		0.0020
		(−0.13)		(0.02)
Constant	−0.9884 ***	−4.4917 ***	0.1799 ***	−0.9110 ***
	(−7.27)	(−6.55)	(4.44)	(−4.22)
Industry Effect	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES
N	9818	9818	9818	9818
R2/Pseudo R2	0.0190	0.0438	0.0230	0.0508

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

show that HSR is still significantly and positively correlated to corporate M&A decisions (M&A) at the 1% significance level, which is consistent with the previous results.

5.5. Control the Impact of Other Transportations

Considering the fact that other transportations can also shorten the distance, we add the other four types of transportation as control variables in the regression. In this section, we manually collect data from the websites of the Civil Aviation Administration of China and the National Bureau of Statistics of China. We choose the natural logarithm of annual passenger throughput (InAir_t) of the prefecture-level administrative district where the listed company is located, the natural logarithm of the total railROAd mileage (InRail_t) of the province where the listed company is located, and the natural logarithm of the total inland waterway mileage (InWater_t) and the total ROAd mileage (InROAd_t) for each year in China as the added control variables.

The regression results obtained using the Probit model are presented in

Table 11. Robustness test (H1): controlling the impact of other transportation.

Variables	M&A
	(1)	(2)	(3)	(4)	(5)
HSR	0.1387 ***	0.1341 ***	0.1354 ***	0.1332 ***	0.1398 ***
	(4.22)	(4.11)	(4.15)	(4.08)	(4.23)
Size	0.2574 ***	0.2563 ***	0.2558 ***	0.2560 ***	0.2561 ***
	(18.29)	(18.26)	(18.19)	(18.22)	(18.23)
ROA	1.1532 ***	1.1553 ***	1.1580 ***	1.1582 ***	1.1531 ***
	(5.99)	(6.00)	(6.02)	(6.02)	(6.00)
Leverage	−0.1165 *	−0.1106 *	−0.1128 *	−0.1115 *	−0.1092 *
	(−1.84)	(−1.75)	(−1.79)	(−1.77)	(−1.73)
Growth	0.0085	0.0083	0.0081	0.0081	0.0087
	(1.21)	(1.19)	(1.14)	(1.15)	(1.23)
OCF	−0.3740 ***	−0.3690 ***	−0.3687 ***	−0.3680 ***	−0.3716 ***
	(−2.82)	(−2.78)	(−2.78)	(−2.77)	(−2.80)
Tobin’ s Q	0.0691 ***	0.0686 ***	0.0685 ***	0.0685 ***	0.0685 ***
	(5.86)	(5.82)	(5.81)	(5.82)	(5.80)
Board	0.0079	0.0107	0.0090	0.0110	0.0080
	(0.11)	(0.15)	(0.12)	(0.15)	(0.11)
Independ	0.1869	0.1907	0.1894	0.1918	0.1888
	(0.83)	(0.85)	(0.84)	(0.85)	(0.84)
Equity	−0.0244	−0.0256	−0.0251	−0.0258	−0.0239
	(−0.32)	(−0.33)	(−0.32)	(−0.33)	(−0.31)
Dual	0.0316	0.0321	0.0321	0.0317	0.0335
	(1.29)	(1.31)	(1.31)	(1.29)	(1.36)
ESH	0.3833 ***	0.3865 ***	0.3812 ***	0.3847 ***	0.3837 ***
	(5.65)	(5.69)	(5.61)	(5.67)	(5.65)
SOE	−0.1822 ***	−0.1851 ***	−0.1849 ***	−0.1856 ***	−0.1834 ***
	(−6.55)	(−6.65)	(−6.64)	(−6.65)	(−6.57)
M/B	−0.0751 ***	−0.0758 ***	−0.0753 ***	−0.0755 ***	−0.0757 ***
	(−6.09)	(−6.15)	(−6.11)	(−6.13)	(−6.14)
PastReturns	20.1744 ***	20.3268 ***	20.1919 ***	20.2583 ***	20.3440 ***
	(9.62)	(9.70)	(9.63)	(9.66)	(9.71)
Volatility	144.0032 ***	143.5880 ***	143.2942 ***	143.5115 ***	143.0560 ***
	(7.13)	(7.10)	(7.09)	(7.10)	(7.08)
LnGDP	0.0295	0.0239	0.0257	0.0247	0.0290
	(1.00)	(0.86)	(0.93)	(0.89)	(0.98)
Open	−0.0284	−0.0536 *	−0.0190	−0.0385	−0.0484
	(−0.96)	(−1.66)	(−0.63)	(−1.26)	(−1.43)
Unemploy	−0.5070 ***	−0.3749 **	−0.4319 ***	−0.4067 **	−0.3319 **
	(−3.28)	(−2.34)	(−2.71)	(−2.53)	(−2.04)
lnAir	−0.0004				−0.0010
	(−0.20)				(−0.55)
lnRail		−0.0342 *			−0.0920 **
		(−1.89)			(−2.06)
lnWater			−0.0041		−0.0078
			(−1.10)		(−1.48)
lnROAd				−0.0166	0.0514
				(−1.32)	(1.44)
Constant	−6.8768 ***	−6.5369 ***	−6.7899 ***	−6.6162 ***	−6.6680 ***
	(−15.26)	(−13.97)	(−15.44)	(−14.08)	(−13.89)
Industry Effect	YES	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899	26,899
Pseudo R2	0.0596	0.0598	0.0597	0.0597	0.0599

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

. The results show that HSR is still significantly and positively related to corporate M&A decisions (M&A) at the 1% significance level, which is consistent with the results of the previous tests. (We also use the fixed effects model and the results are similar. For brevity, we do not show the results of the fixed effects model below).

The regression results obtained using the Probit model are presented in

Table 12. Robustness test (H2): controlling the impact of other transportation.

Variables	M&A
	(1)	(2)	(3)	(4)	(5)
ChenJi	−0.0255	−0.0268	−0.0287	−0.0302	−0.0184
	(−0.48)	(−0.50)	(−0.53)	(−0.57)	(−0.34)
NonChenJi	0.1162 ***	0.1108 ***	0.1140 ***	0.1119 ***	0.1170 ***
	(3.58)	(3.43)	(3.54)	(3.47)	(3.59)
TwoTrain	0.1344 **	0.1310 **	0.1336 **	0.1332 **	0.1270 **
	(2.40)	(2.34)	(2.38)	(2.38)	(2.26)
Size	0.2560 ***	0.2550 ***	0.2555 ***	0.2550 ***	0.2559 ***
	(18.19)	(18.16)	(18.15)	(18.15)	(18.18)
ROA	1.1393 ***	1.1410 ***	1.1399 ***	1.1421 ***	1.1358 ***
	(5.91)	(5.92)	(5.91)	(5.93)	(5.89)
Leverage	−0.1121 *	−0.1076 *	−0.1117 *	−0.1095 *	−0.1078 *
	(−1.78)	(−1.71)	(−1.77)	(−1.74)	(−1.71)
Growth	0.0082	0.0080	0.0080	0.0079	0.0086
	(1.17)	(1.13)	(1.13)	(1.12)	(1.22)
OCF	−0.3530 ***	−0.3480 ***	−0.3502 ***	−0.3483 ***	−0.3537 ***
	(−2.65)	(−2.62)	(−2.64)	(−2.62)	(−2.66)
Tobin’ s Q	0.0685 ***	0.0680 ***	0.0683 ***	0.0681 ***	0.0683 ***
	(5.81)	(5.77)	(5.79)	(5.78)	(5.79)
Board	0.0061	0.0083	0.0063	0.0080	0.0054
	(0.08)	(0.11)	(0.09)	(0.11)	(0.07)
Independ	0.1648	0.1674	0.1642	0.1673	0.1643
	(0.73)	(0.75)	(0.73)	(0.75)	(0.73)
Equity	−0.0210	−0.0223	−0.0217	−0.0224	−0.0201
	(−0.27)	(−0.29)	(−0.28)	(−0.29)	(−0.26)
Dual	0.0324	0.0330	0.0326	0.0326	0.0339
	(1.32)	(1.34)	(1.33)	(1.33)	(1.38)
ESH	0.3766 ***	0.3794 ***	0.3763 ***	0.3778 ***	0.3785 ***
	(5.55)	(5.59)	(5.54)	(5.56)	(5.58)
SOE	−0.1881 ***	−0.1908 ***	−0.1892 ***	−0.1905 ***	−0.1877 ***
	(−6.76)	(−6.85)	(−6.80)	(−6.82)	(−6.73)
M/B	−0.0749 ***	−0.0756 ***	−0.0751 ***	−0.0753 ***	−0.0754 ***
	(−6.07)	(−6.13)	(−6.09)	(−6.10)	(−6.11)
PastReturns	19.9912 ***	20.1294 ***	20.0073 ***	20.0550 ***	20.1496 ***
	(9.53)	(9.60)	(9.54)	(9.56)	(9.61)
Volatility	144.6036 ***	144.2538 ***	144.4646 ***	144.3170 ***	144.2091 ***
	(7.16)	(7.14)	(7.15)	(7.14)	(7.13)
LnGDP	0.0216	0.0137	0.0161	0.0152	0.0215
	(0.72)	(0.49)	(0.57)	(0.54)	(0.71)
Open	−0.0111	−0.0317	−0.0090	−0.0177	−0.0366
	(−0.37)	(−0.98)	(−0.30)	(−0.57)	(−1.08)
Unemploy	−0.3886 **	−0.2795 *	−0.3663 **	−0.3248 **	−0.2759 *
	(−2.51)	(−1.74)	(−2.28)	(−2.01)	(−1.69)
lnAir	−0.0010				−0.0015
	(−0.56)				(−0.78)
lnRail		−0.0278			−0.0852 *
		(−1.53)			(−1.91)
lnWater			−0.0009		−0.0041
			(−0.23)		(−0.75)
lnROAd				−0.0102	0.0469
				(−0.80)	(1.33)
Constant	−6.7429 ***	−6.4393 ***	−6.6831 ***	−6.5546 ***	−6.5832 ***
	(−14.92)	(−13.75)	(−15.15)	(−13.97)	(−13.68)
Industry Effect	YES	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES	YES
N	26,899	26,899	26,899	26,899	26,899
Pseudo R2	0.0604	0.0605	0.0604	0.0605	0.0607

Definitions of variables are presented in Table 1. ***, **, and * indicate 1%, 5%, and 10% significant, respectively.

. The results show that NonChenJi and TwoTrain are still significantly and positively related to corporate M&A decisions (M&A) at the 1% significance level, which is consistent with the results of the main tests. (We also use the fixed effects model and the results is similar. For brevity, we do not show the results of the fixed effects model below).

6. Conclusions

Using the DID model and a sample of A-share listed companies from 2006 to 2019, this paper examines the impact of the quasi-natural experimental event of HSR opening on enterprises’ M&A behavior. We find that the opening of HSR can promote the M&A activities of listed companies in the corresponding prefecture-level administrative districts, and the opening of non-intercity HSR has a more obvious effect on the promotion of M&A decisions of enterprises. These results remain true after a series of robustness tests.

Taking the opening of HSR as an entry point, we explore the impact of objective environmental changes on operating decisions. This paper adds new research perspectives and empirical evidence on the economic consequences of HSR opening as well as the factors influencing corporate M&A behavior, and expands the scope of application of the new economic geography at the micro level.

Finally, the prospects for future research should be mentioned here. With the outbreak and spread of the COVID-19 pandemic, travel restriction has decreased people’s willingness and opportunities to take public transportation (including HSR), which may influence the transportation economics (Donthu and Gustafsson, 2020) [36]. The existing literature has started to explore the consequences of COVID-19 changes at the macro level, such as economic development and capital market. However, given the limited number of papers on the topic of the COVID-19 pandemic and transportation economics, our understanding is still inadequate and superficial (Lim, 2021) [37]. Therefore, we will seek to add to our knowledge about the micro-level consequences of changes (e.g., at the individual and firm levels) to help us better understand economics and business management in a fast-changing world.