The Impact of the Belt and Road Initiative on the Capital Market of Chinese Listed Companies: From the Perspective of Sustainable Development

Stock liquidity is recognized as a central component of the capital market [15]. A substantial body of research has been devoted to understanding the factors that influence stock liquidity. At the macro level, studies have examined the effects of monetary liquidity [16, 17], exchange rate policy [18], and monetary policy on stock liquidity [19].

Conversely, at the micro level, research has focused on the impact of factors such as foreign ownership [20], stock market activity [21], overseas direct investment [22], and risk information disclosure on stock liquidity [23].

Interestingly, no study to date has explored the influence of the B&R Initiative on stock liquidity. However, it is reasonable to infer that the impact of the B&R Initiative on corporate capital market performance is closely intertwined with the aforementioned factors shaping stock liquidity.

Firstly, it is noted that the Chinese government plays a pivotal role in guiding direct investment along the Belt and Road Initiative (B&R) [24]. In the B&R context, corporations investing in B&R countries are likely to attract more investor and media attention compared to those investing in non-B&R countries, potentially leading to improved stock liquidity [25, 26].

In accordance with Johanson's theory of outsider disadvantage [27], B&R countries, having signed relevant cooperation agreements with China, pose less discriminatory harm due to legitimacy deficits and relationship harm from a lack of embeddedness for Chinese enterprises. As such, the additional costs of direct investment in B&R countries are likely to be lower than those in non-B&R countries. These cost reductions, when reflected in the capital market, may enhance stock liquidity.

The theory of empathy effect suggests that cultural institutional distance can influence the overseas communication costs of investment enterprises. Compared to non-B&R countries [28, 29], China and Asian B&R countries demonstrate closer ties across most dimensions, having a long history of trade, cooperation, and a higher foundation of trust [30]. An increase in manager's perception of host country similarity could encourage corporate social responsibility fulfillment [31].

There are, however, differing views regarding the relationship between social responsibility and stock liquidity. Some argue that social responsibility information increases analyst attention and reduces corporate stock liquidity [32], while others posit that a higher degree of social responsibility fulfillment improves stock liquidity [33]. Capital market investors may recognize the social responsibility of enterprises by increasing stock liquidity [34].

Moreover, overseas direct investment's impact on corporate innovation is influenced by the heterogeneity of the host countries. Notably, over 80% of B&R countries are developing nations [35]. Scholars suggest that when enterprises invest in developed countries, it aids in information disclosure [36], which, in turn, could influence stock liquidity. For risk information disclosure, a negative correlation with stock liquidity is observed, suggesting that increased risk disclosure leads to stronger market risk perception, more cautious investor behavior, and lower stock liquidity [23]. Contrastingly, from the perspective of stock bar activity, higher activity can aid enterprises in disclosing useful information, thereby reducing uninformed traders' adverse selection and potentially leading to higher stock liquidity [21].

H_1a: The direct investment of B&R countries will increase the stock liquidity compared with that of non B&R countries.

H_1b: The direct investment of B&R countries will reduce the stock liquidity compared with that of non B&R countries.

The attention garnered by overseas investment events has been observed to gradually diminish over time, subsequently leading to a decrease in investor and media focus [22]. As the period of overseas investment extends, enterprises engaged in foreign direct investment will likely encounter a reduction in impact, as postulated by both the theory of outsider disadvantage and the theory of empathy effect. Considering this dynamic effect, the following research hypothesis is proposed:

H₂: The impact of the direct investment of B&R countries on the liquidity of corporate stocks will weaken over time.

4.1 Descriptive statistics

To ensure consistency between the samples used in descriptive statistics and those entering the regression, we showed the descriptive statistical results of the variables of this article in Table 1. The mean of the illiquidity indicator (ILLIQ_Y) is 0.071, the maximum value is 2.341, and the minimum value is 0.0016, indicating a significant gap in the illiquidity indicators in the study sample. The main descriptive statistical results are as follows: the mean of enterprise size is 22.6791, the mean of equity concentration (SD) is 34.6489, and the mean of return on equity (ROE) is 0.0675, which is basically consistent with Wu et al. [38]. Among them, the mean of age is 9.5411, and the mean of book to market ratio (BM) is 0.6359, which is significantly different from the study by Wu et al. [38]. The reason is that in terms of sample selection, this article selects listed companies with overseas direct investment, which has significant differences in terms of age and market value compared to the entire sample of companies.

Table 1. Descriptive statistical results

4.2 Benchmark regression results

Table 2 presents the regression analysis results delineating the influence of the Belt and Road (B&R) initiative on corporate stock liquidity. The estimations provided in column (1) incorporate firm and year-fixed effects, while column (2) extends the model by including a set of control variables. Across all models, the regression coefficient for the B&R variable is consistently negative, suggesting that the implementation of the B&R initiative has served to enhance corporate stock liquidity. This outcome provides empirical support for $H_{1 \mathrm{a}}$.

Table 2. Benchmark regression results

Note: (1) * * *, * *, * respectively represent significance levels at 1%, 5%, and 10%; (2) The data in parentheses is the t-value adjusted by the robust standard of clustering. The following table is the same.

4.3 Heterogeneity test

Considering the actual situation in China, under different enterprise nature conditions, the impact of the B&R on stock liquidity may have asymmetric effects. Based on this, this article divides the entire sample into state-owned enterprises and non-state-owned enterprises referring to their attributes. In the group of state-owned enterprises, the regression coefficient of the B&R on stock liquidity is positive and has not passed the statistical significance test. The regression coefficient of the B&R on stock flow is negative and has passed the 5% statistical significance test in the state-owned enterprise. The heterogeneity test table of non-state-owned enterprise and non-state-owned enterprise is shown in Table 3. The reasons for the heterogeneity between state-owned and other enterprises may be as follows: the former are the executors of national policies and have "non market motives" [40]. Therefore, the state-owned enterprises' direct investment in R&D countries will not significantly change the attention of their investors and media than the non-state-owned enterprises. The investment of the former enterprises in R&D countries mostly flows to high-risk countries [40]. The reason why the liquidity of state-owned enterprise stocks has not significantly decreased is analyzed as follows. When state-owned enterprises disclose risks, they often focus more on macro policy risks, while the heterogeneity risk of enterprises is relatively weak, which to some extent diverts investors' attention from enterprise risks [23].

Table 3. Heterogeneity test table

4.4 Parallel trend test

Before adopting the DID model, it should be ensured that the experimental and control group have a consistent trend when the policy is changed [41]. Due to the different time of enterprises' foreign investment in the B&R countries, a virtual variable of time cannot be set just by taking a certain year as the breakthrough point of policy impact. In fact, we should set relative dummy variables for the B&R enterprises. In this paper we used the research method of Bai [41] to construct the following formula for parallel trend test and dynamic effect analysis:

$\begin{aligned} & \text { ILLIQ_Y}_{i, t}=\alpha+\beta_1 \text { Before 2}_{i, t}+\beta_2 \text { Before } 1_{i, t} +\beta_3 \text { Current }_{i, t}+\beta_4 \text { After 1}_{i, t} +\beta_5 \text { After 2}_{i, t}+\gamma \text { Control }_{i, t} +F i r m F E+Y e a r F E+\varepsilon_{i, t}  \end{aligned}$                      (3)     

Among them, the time dummy variable is the observation value of the first n years, this year and the next n years of the B&R enterprises' investment in the B&R countries. The dummy variables of non B&R enterprises are all 0. Given the observation period spanning from 2013 to 2022, both firm and year-fixed effects were controlled for during the regression analysis. The outcomes of the regression are graphically presented in Figure 1, where the solid point denotes the regression coefficient and the bounding solid lines represent the 95% confidence interval, thereby providing a visual interpretation of the statistical significance and uncertainty of the estimated effects.

pi_zhu_2023-10-30_210517.png

Figure 1. Parallel trend test chart

From the results we could find out that before investment in the B&R countries, the relative time dummy variable coefficients are not remarkable, and the values are in the low level. So, we could make a conclusion there exist no significant difference in stock liquidity between the experimental and control group before investing the B&R countries, which conforms to the hypothesis of parallel trend. In the year of investment in the B&R countries, the impact coefficient of direct investment in the B&R countries on stock liquidity is significantly negative. In the year of investment in the B&R countries, the impact coefficient of investment in the B&R countries on stock liquidity is still significantly negative. But compared to that year, its regression coefficient has increased, indicating that the impact has weakened compared to that year. The impact coefficient of the B&R countries' investment on stock liquidity is no longer significant one year after investing in the B&R countries. This result shows that the impact of investment in the B&R countries on stock liquidity is not long-term, so $\mathrm{H}_2$ has been validated.

4.5 Placebo test

Although this article has adopted DID in model design and controlled for firm fixed and year fixed effects to minimize the impact of endogeneity. And in regression equations, we have applied the t-statistic of Cluster robust standard error adjustment improve the robustness of the model regression. However, there is still a possibility that other policies and random factors may have an impact on stock liquidity. To further mitigate the adverse effects of this on the core research results of this article, this article will draw on the research method of Bai [41]and use a placebo test method. Construct 1000 random impacts of pseudo B&R national investment on 432 sample enterprises, and randomly select 87 enterprises as the experimental group each time, and give the time of B&R national investment at random to obtain 1000 sets of dummy variables $B R^*\left(\right.$ Group $^* \times$ Post $\left.^*\right)$; Add 1000 $\beta^*$ kernel density and its p-value distribution is shown in the figure. In Figure 2, the X-axis is the estimated coefficient of $B R^*$ generated 1000 times randomly. The empty circle symbol is the p value of the estimated coefficient, the solid line is the kernel density distribution, and the left vertical line is the estimated coefficient of the actual regression -0.0757. The results show that the random processing generated $\beta^*$ is mainly close to 0, and the P-value is mostly higher than 0.1, with only a few regression P-values smaller than the true regression coefficient. This also indicates to some extent that the core result of the article has a certain degree of robustness.

image020.png

Figure 2. Placebo test chart