Impacts of the Belt and Road Initiative on regional outward FDI from China based on evidence from 2000 to 2015

2.1 Location choices of Chinese MNEs

Although China and its trade partners are mutually dependent, inevitably, there is a power imbalance (or asymmetry) among the actors owing to the uneven distribution of resources or different stages of development (Emerson, 1962; Xia et al., 2014). Usually, there are two strategies the weaker party can adopt to achieve successful resource exchange: adaptation and avoidance (Salancik & Pfeffer, 1978). Adaptation, which also can be referred to as compliance, can be problematic, however, as a weaker party can lose its decision-making autonomy to the stronger party (Nienhüser, 2008; Oliver, 1991). Therefore, diversification (or avoidance) becomes a more promising strategy as it allows the weaker party to still maintain an exchange relationship while tilting the power imbalance. Diversification strategy refers to when an actor in a relatively weak position intends to escape or diversify to other exchange relationships to reduce the constraining influence of the dominant actor (Xia, 2010; Pfeffer, 1987).

The diversification logic is applied in this study to explain Chinese firms’ intention to enter BRI countries for global expansion. As mentioned above, China’s economic development over the last four decades has relied significantly on foreign MNEs operating in China, with the most inward FDI coming from advanced economies such as the U.S. and Western Europe. These foreign MNEs are powerful actors (Hoskisson et al., 2000) because of technology and managerial skill advancement (Inkpen & Beamish, 1997; Yan & Gray, 1994). Although foreign MNEs in China transfer knowledge to local firms, they also have imposed constraints on local firms regarding turn-around time, delivery standards, pricing, etc. (Liu et al., 2009). These constraints overload Chinese suppliers with production orders under stringent quality and efficiency requirements (Guler et al., 2002), eventually locking Chinese manufacturers into a vicious low value-added activity cycle (Gill & Kharas, 2015; Islam & Chadee, 2021). In addition, foreign MNEs dominate major consumer markets, influencing customer tastes and service standards (Luo & Park, 2001).

Furthermore, foreign MNEs have driven up factor prices such as wages (Girma et al., 2019) and have established technology standards that do not favor local firms (Lee, 2005). As a result, Chinese enterprises have found it difficult to prosper or survive under these constraints. The over-dependence on the western developed world thus has caused Chinese firms to seek growth opportunities elsewhere, especially in countries where advanced economy firms have a low influence (Driffield, 2006).

The BRI policy fits the needs of Chinese MNEs seeking non-Western alternative business partners. By building infrastructure in neighboring countries and strengthening diplomatic relations, Chinese MNEs potentially find it easier to have business trips and search for collaborators and opportunities in the BRI countries. In addition, the BRI countries overall are less influenced by advanced economies and are geographically more proximate to China. Therefore, the BRI opens another avenue for Chinese firms to go global. This especially is the case for firms in an unfavorable power position with advanced economy firms at home, as BRI opportunities reduce their dependence on them for markets and technology and, thus, increase their relative power by diversifying into new geographic locations. The above arguments, then, lead to the first hypothesis:

H1: A Chinese firm that originates from a province with more inward FDI from Western Europe and North America is more likely than a firm from other provinces with less such FDI to enter BRI participating countries for its FDI activities after the launch of the BRI.

2.2 FDI motives

Even though the BRI at this stage consists of a large amount of infrastructure building in the host countries, when it comes to FDI activities in BRI regions, Chinese firms also leverage the BRI to expand their other business activities, such as market expansion and manufacturing, which are congruent with the all-inclusive goal of the BRI as stated in its official document. These diversified business motives include trade-supportive investment, market-seeking FDI, efficiency-seeking FDI, and strategic asset-seeking FDI.

According to the China Statistical Yearbook 2019, 41.66 % of China’s exports are conducted by foreign MNEs in China. When Chinese domestic firms partner with these foreign MNEs at home, their contractual agreements usually indicate that they are not supposed to export to the foreign MNEs’ target markets. This prevention of potential competition from Chinese firms by foreign MNEs impedes the growth of Chinese firms’ international trade. Nevertheless, they do have a strong intention to promote and facilitate the export and import of goods and services. With the launch of the BRI, Chinese firms can be expected to be eager to take advantage of the opportunity and expand to BRI countries for trade-supportive investment. This, therefore, leads to the following hypothesis:

H2a: A Chinese firm that originated from a province with more inward FDI from Western Europe and North America, compared to a firm from other provinces with less such FDI, is more likely to enter BRI countries for trade-supportive FDI after the launch of the BRI.

In the mid-1980s, the Chinese government promoted international joint venture entry modes with the notion that the country needed to give up portions of the domestic market in exchange for advanced technologies (Wei & Davis, 2018). Since a joint venture is considered an MNE subsidiary, the joint venture partners usually receive in-house technology transfer and employee training opportunities (Li & Cantwell, 2012). While Chinese firms’ manufacturing capacity has improved thanks to foreign knowledge transfer, the fast expansion of manufacturing capacity has also led to over-capacity issues in China. Also, owing to foreign competition at home, Chinese firms’ market share has been affected. With the launch of the BRI, however, Chinese firms can be expected to explore alternative markets, hoping to transfer their overcapacity to and increase their market share in countries with more favorable diplomatic relations and similar cultural and social norms. Hypothesis 2b follows:

H2b: A Chinese firm that originates from a province with more inward FDI from Western Europe and North America, compared to a firm from other provinces with less such FDI, is more likely to enter BRI countries for market-seeking FDI after the launch of the BRI.

Foreign MNEs in China have driven up factor prices, including wages and other manufacturing costs. The FDI spillover literature (e. g., Blomstrom & Kokko, 2001) has documented the employment effect. Foreign firms tend to pay higher wages to attract local talent upon entry. Nevertheless, when workers from foreign firms become boundary spanners, switch to local companies or become entrepreneurs, their expectations for wages and standards of living rise (Girma et al., 2019). Foreign MNEs’ standards in workplace safety and employee welfare, including health insurance and retirement contributions, also urge local companies to converge on global practices. After the Foxconn scandal of forced labor, let alone the civil-level movement on improving working conditions, the Chinese government raised the minimum wage by 9 % (Hoffman, 2014). As a result, it gradually has become difficult for Chinese firms to keep their competitive edge by producing at home. Manufacturing costs even increase when Chinese firms co-locate with foreign MNEs, as foreign MNEs’ adjacent areas are likely to experience wage increases (Xia et al., 2014). Chinese firms seldom have found that locations in advanced economies solve this problem – owing to the high labor costs and rents in these economies – even if Chinese firms established business networks there (Hertenstein et al., 2017). Therefore, the BRI provides essential opportunities for Chinese firms to shift their manufacturing activities, which also benefits them in their industrial upgrading efforts. This, then, leads to the next hypothesis:

H2c: A Chinese firm that originated from a province with more inward FDI from Western Europe and North America, compared to a firm from other provinces with less such FDI, is more likely to enter BRI countries for efficiency-seeking FDI after the launch of the BRI.

Although the Chinese government intended to temporarily give up the domestic market in exchange for advanced technologies, things have not always occurred as expected. Previous research has shown that cross-border technology transfer, whether in-house or internal, can be ill-functioning and not achieve its goal (Inkpen, 2000; Buckley et al., 2004). Also, foreign MNEs in weaker intellectual property regimes tend to purposely avoid sharing cutting-edge technologies (Zhao, 2006). Nevertheless, Chinese firms with interdependent relationships and intensive interactions with foreign MNEs perceive a large technology gap with foreign MNEs. Therefore, these Chinese firms are eager to upgrade their technological capabilities and foster domestic and international collaborations. Since advanced economies are becoming more cautious concerning Chinese firms learning advanced technologies from them and are concerned about their image in doing this (Cuervo-Cazurra et al., 2014), the BRI has enabled Chinese enterprises to now seek knowledge in BRI countries as an alternative. In Singapore, one of the BRI destination countries, strategic-asset seeking FDI projects are very similar to those in the U.S. and Western Europe, covering a broad spectrum of pharmaceutical research, medical devices, computer hardware and software, automobile, and machinery development, etc. In Russia, strategic asset-seeking FDI projects mainly focus on heavy-duty machinery, such as mining equipment. In India, half of the strategic asset-seeking projects are in the computer software domain. Thus, the last hypothesis, which focuses on strategic asset-seeking FDI, is:

H2d: A Chinese firm that originates from a province with more inward FDI from Western Europe and North America, compared to a firm from other provinces with less such FDI, is more likely to enter BRI countries for strategic asset-seeking FDI after the launch of the BRI.

3.1 Data sample

The Outward FDI Directory (OFDI Directory) published by the Ministry of Commerce of China was used to test the above hypotheses. The current OFDI directory includes 41.707 projects between 1983 and 2015, providing adequate observations to compare before and after the BRI launch (at least for its initial years). The OFDI Directory has proven to be a legitimate source for Chinese project-level outward FDI studies (Yang & Bathelt, 2021). The OFDI Directory covers 31 Chinese provinces and 201 host destinations and documents non-financial sector project-level outward FDI activities from China. The dataset offers the following information: parent firm name, subsidiary name, OFDI date, host country, home province, and manager self-reported subsidiary activities. After removing tax haven cases, 27.547 are left, with 11.124 going to BRI countries between 2000 and 2015 (Table 1).

3.2 Variables and measurements

For hypothesis 1, the dependent variable is BRI destination. The value is coded as 1 if the outward FDI project entered one of the BRI countries and 0 otherwise. Logistic regression is used to calculate the likelihood of an outward FDI in one of the original 64 BRI countries.

The dependent variable for hypotheses 2a, 2b, 2c, and 2d, which refer to BRI investment only, is FDI motive. Based on Dunning (1993) and Cuervo-Cazurra & Narula (2015), five major motives are associated with Chinese outward FDI projects: natural resource-seeking; trade-supportive; market-seeking; efficiency-seeking; and strategic asset-seeking. Because of the particularity of the BRI, a sixth motive, infrastructure building, can be added. The hypotheses tested in this paper address those four motives where the impact is less clear. What is worth mentioning is that, unlike for trade-supportive investment in which production is home-based, MNEs have procurement or sales offices overseas, market-seeking FDI usually incurs local production and local sales abroad (Johanson & Vahlne, 1977).

The motives are coded based on the managers’ self-reported activities in the host country. These six motives are listed as separate dependent variables (Table 2). For example, if one project has a market-seeking motive, then “market-seeking” as a dependent variable will be coded as 1, otherwise 0. It should be noted that one project can have multiple FDI motives.

The empirical models of this research use two predictor variables, time and treated.

Time: Chinese President Xi Jinping unveiled the initiative in September and October 2013 during his visits to Kazakhstan and Indonesia. These visits signaled the launch of the BRI and prepared Chinese firms to start their BRI journey. As a result, 2013 is used as the cut-off year to proxy the policy year; years after 2013 are coded as 1, and years before 2013 are coded as 0.

Treated: A province with a large volume of inward FDI stock, namely provinces on the east coast, is coded as 1, whereas western provinces are coded as 0. The east coast provinces in China include Anhui, Henan, Shandong, Shanghai, Beijing, Hunan, Jiangsu, Zhejiang, Guangdong, Fujian, Hebei, Sichuan, Jiangxi, Tianjin, Shanxi, Hubei, Guizhou, and Hainan.

State ownership (SOE) is used as a control, as many state-owned enterprises participate in and promote the BRI (Buckley, 2020; Li & Zeng, 2019; Sutherland et al., 2020).

3.3 Empirical technique

This paper applies maximum likelihood logistic regressions since the dependent variables are dichotomous. The identification strategy is based on a difference-in-difference (DID) analysis that compares the likelihood of location choices and motive choices before and after the launch of the BRI program. 2013 is recognized as the cut-off point in launching the BRI policy by multiple studies (e. g., Beule & Zhang, 2022; Buckley, 2020; Lewin & Witt, 2022). In this study, the first difference refers to the different types of provinces, namely provinces with more FDI from Western Europe and North America (eastern Chinese provinces) and provinces with fewer such FDI (western Chinese provinces). The second difference compares periods before and after the launch of the BRI policy. The two differences are simultaneously considered in the full model.

4.1 Summary statistics

The following section presents summary statistics on the different samples and variables before the results and findings are discussed. The number of FDI projects grew significantly in 2014 and 2015 after the launch of the BRI, equaling 52 % of the total number of projects between 2000 and 2013. Regarding the distribution of Chinese outward FDI in the 64 BRI regions, East Asia attracts half of the investment projects. Although Central and East Europe (CEE) has the largest number of countries, it has the lowest number of Chinese FDI projects.

For statistics at the province level, 13 western provinces are BRI-participating provinces. Nevertheless, most of the investments in BRI countries come from eastern provinces, with 73.3 % of the total investment project count in BRI countries. The time period 2014–2015 accounts for one-third of the same total investment.

In addition, inward FDI in each province was investigated. 18 non-BRI participating provinces, also known as eastern provinces, have accumulated 7.5 times the inward FDI by dollar amount than western provinces. Therefore, a province’s location in the East or West is used as a proxy for its reliance on inward FDI which is mainly from Western Europe and North America.

Besides the host country and home region, FDI motives also are crucial in this analysis. Out of 11.114 projects between 2000 and 2014 in BRI regions, 982 conducted infrastructure-building activities, and 1685 had natural resource-seeking components. These two are commonly-noticed FDI activities in BRI regions and will not be tested in this paper. Nevertheless, market-seeking FDI is the most frequent investment case, with 6845 projects. Trade-supportive investment and efficiency-seeking FDI are ranked next in project counts, with both in the range of 4000. Unsurprisingly, strategic asset-seeking FDI is the least frequent, with only 658 cases. However, strategic asset-seeking FDI, along with infrastructure-building projects, has seen the fastest growth, with their number in the most recent two years almost the same as in the previous two decades. It should be noted, too, that since one FDI project can have multiple FDI motives, the number of individual motives adds up to more than the total project count.

4.2 Regression results

This section discloses the regression results. Tables 3 through 7 present the empirical results in terms of hypothesis testing. For hypothesis 1, the larger 27.547 sample, which contains both BRI and non-BRI countries, was used. The correlation among variables is between –0.2 and 0.15, and all VIF values are below 2, indicating that there are no multicollinearity concerns. Model 1 and Model 2 in each table show the main effects, while model 3 includes the difference-in-difference moderating effect. The first difference time (0 or 1) refers to FDI entry timing before or after the launch of the BRI. The second difference treated (0 or 1) compares east and west provinces.

Table 3 focuses on location choices, showing the logit regression results in testing hypothesis H1. The results in this table aim to explain whether an outward FDI project will be located in a BRI country (or not). Province, east or west, and outward FDI timing are predictors while controlling for state ownership. After the launch of the BRI (time=1), the coefficient of time being negative and significant shows that in the early years after the launch of the BRI policy, Chinese firms are still less likely to switch to BRI regions for their FDI destinations. The results also show that eastern provinces^[1] (treated=1) are less likely to invest in BRI countries than western provinces, as the coefficient for treated is negative and significant across all models.

Furthermore, the difference-in-difference (time*treated=1) effect explains the location choice of firms from eastern provinces after the launch of the BRI. The coefficient of time*treated being negative and non-significant shows that eastern provinces do not exhibit a significant difference in their tendency to invest in BRI countries before or after the policy launch. As a result, hypothesis H1 is not supported.

The BRI sample, which contains only those 11.124 FDI projects in BRI regions, is used to test hypotheses 2a, 2b, 2c, and 2d. The correlation coefficient among all variables is between –0.3 and 0.3. In addition, the VIF values are below 2.5 for each regression. Therefore, there are no multicollinearity concerns.

Table 4 displays the tests for hypothesis H2a, whether the BRI promotes trade-supportive investments. The positive and significant coefficient of the variable time across all models in Table 4 shows that the likelihood of an FDI project with a trade-supportive motive is higher after the launch of the BRI (time=1). In addition, firms from eastern provinces (treated=1) have a lower propensity than firms from western provinces to initiate a trade-supportive FDI project. In other words, trade-supportive investment in BRI regions is more likely to originate from western provinces. However, the insignificant coefficient of the difference-in-difference term indicates that the east or the west provinces’ tendencies in engaging in trade-supportive investment do not change before or after the BRI was launched. Therefore, the results in table 4 do not support hypothesis H2a.

Table 5 explains the market-seeking tendencies in hypothesis H2b. The positive and significant coefficient of predictor time indicates that, generally, firms are less likely to have market-seeking FDI in the BRI regions after the launch of the BRI (time=1). In addition, it can be seen from the positive and significant coefficient of the predictor treated that firms from eastern provinces (treated=1) are more likely than those from western provinces to engage in market-seeking activities in BRI regions.

The difference-in-difference term (time*treated) is significant but negative in Table 5. Based on the equation Marketing Seeking = α+ β₁ SOE + β₂ Time + β₃ Treated + β₄ Time*Treated + γ, scenario (1) treated = 1 and time = 0 and scenario (2) treated = 1 and time = 1 can be compared. The first scenario represents the likelihood of conducting a market-seeking FDI for eastern province firms before the launch of the BRI, when treated = 1 and time = 0, Marketing Seeking = α- 0.061 * SOE + 0 + 0.150 + 0 + γ = α+ 0.150 –0.061 SOE + γ. In the second scenario, when treated = 1 and time = 1, the equation reads Marketing Seeking = α – 0.061 * SOE – 0.141 + 0.150 –0.161 + γ = α – 0.152 –0.061 SOE + γ. The difference between scenario (1) and scenario (2) is 0.150–0.152= – 0.002. This negative value indicates that marketing-seeking FDI is more likely to happen in scenario (2) than in scenario (1). In other words, the likelihood for eastern province firms to enter BRI countries for market-seeking projects is higher after the launch of the BRI. Therefore, the market-seeking motive results support hypothesis H2b.

Table 6 displays the test of the efficiency-seeking FDI hypothesis H2c. The coefficient of predictor time is positive and significant in the main effect models, suggesting that, in general, all firms are more likely to have efficiency-seeking FDI in BRI regions after the BRI’s launch (time=1). The results also indicate that firms from eastern provinces (treated=1) are less likely to engage in efficiency-seeking activities in BRI regions than those from western provinces.

The difference-in-difference term in model 3 is positive and significant in Table 6. The difference-in-difference method compares the likelihood of efficiency-seeking FDI before and after the launch of BRI. When treated = 1 and time = 0, the likelihood for efficiency-seeking is –0.632 – 0.672 soe. After the launch of the BRI, when treated = 1 and time = 1, the likelihood of efficiency-seeking is – 0.632 – 0.240 + 0.541 – 0.672 soe. Therefore, the likelihood in scenario two is larger than that in scenario one. This result indicates that firms from eastern provinces, compared to firms from western provinces, are more likely to enter BRI countries for efficiency-seeking FDI projects after the launch of the BRI. Therefore, the results support hypothesis H2c such that firms from eastern China, where firms are more likely to be subject to the influences of advanced country MNEs, are more inclined to seek cost-effective alternative destinations for manufacturing than firms from western China.

Table 7 presents the results of the strategic asset-seeking hypothesis H2d. The predictor time being positive and significant in the main effect models indicates that, in general, firms are more likely to have strategic asset-seeking FDI in BRI regions after the BRI was launched (time=1). We can also learn from the coefficient of variable treated in Model 1 and Model 2 that, generally, firms from eastern provinces (treated=1) are more likely than firms from western provinces to engage in strategic asset-seeking activities in BRI regions.

The difference-in-difference term treated*time is negative and significant in the logit regression in Table 7. The following two scenarios are compared to determine how the BRI policy has affected firms from eastern provinces. The first scenario shows that when treated = 1 and time = 0, the likelihood for strategic asset-seeking FDI conducted by eastern province firms is 0.053 + 0.129 soe. In the second scenario, when treated = 1 and time = 1, the likelihood of strategic asset-seeking is 0.053 + 0.921 – 0.438 + 0.129 soe. Therefore, the likelihood in scenario two is larger than in scenario one. This result indicates that eastern provinces are more likely to enter BRI countries for strategic-asset seeking FDI projects after the BRI was initiated. After the BRI began, eastern province firms have become even more eager to seek technology cooperation and local talent in their FDI projects than western province firms. Therefore, the results in table 7 provide support for hypothesis H2d. Firms from eastern China, subject to much higher inward investments from the advanced countries, are more likely to seek alternative host countries for knowledge-seeking activities than firms in western China.

To conduct a robustness check on the FDI motive analysis, the country profile of the two supported hypotheses was explored after the regression results. For efficiency-seeking FDI, Russia and other East Asian countries rank top on the list^[2]. According to the data, most Chinese firms entering Russia for efficiency-seeking FDI are from the northeastern part of China: Heilongjiang Province, Jilin Province, and Liaoning Province. These three provinces belong to BRI participating provinces. The efficiency-seeking projects in Russia mainly include processing agricultural and dairy products, and wood furniture from lumberyards. Nevertheless, efficiency-seeking projects in East Asia are in the manufacturing sector, such as textiles and garments, rubber and plastic, white goods, etc. Unlike in Russia, most Chinese firms in East Asia for efficiency-seeking FDI originate from non-BRI provinces.

For strategic asset-seeking FDI, Singapore, Russia, and India dominate the top 10 list^[3]. The data reflects that strategic asset-seeking projects are mainly from eastern provinces, known as non-BRI provinces. In Singapore, strategic-asset seeking FDI projects are similar to those in the U.S. and western Europe, covering a broad spectrum including pharmaceutical research, medical devices, computer hardware and software, automobile, machinery development, etc. In Russia, strategic asset-seeking FDI projects mainly focus on heavy-duty machinery, including mining equipment, cranes, logging trucks, etc. In India, half of the strategic-asset seeking projects are in the computer software domain.

The summary of the results and findings suggests that the signs and the significance levels of single explanatory variables time and treated largely converge to expectations. In terms of time, Chinese FDI in BRI regions was expected to increase after the policy launch. The results show that Chinese FDI projects in trade-supportive and strategic asset-seeking motives have increased in BRI regions after the policy’s launch, but market-seeking activities and efficiency-seeking do not show the same trend. This finding is consistent with BRI regions’ relatively low GDP per capita, indicating the regions’ overall low purchasing power and less competitive manufacturing capacities. Regarding the variable treated, which differentiates eastern or western provinces in China, it is expected that firms from eastern provinces are more likely to conduct FDI activities that require higher-level internationalization experiences and capabilities. The results comply with the expectations that firms from eastern part of China are indeed more likely to engage in trade-supportive investments and market-seeking and strategic asset-seeking FDI.

The results for the difference-in-difference term do not support the location choice hypothesis but largely support the FDI motives hypotheses. H1 and H2a are not supported, while H2b, H2c and H2d are. H1 expects that the BRI will lead Chinese firms from eastern provinces to invest in BRI regions. Nevertheless, the results do not show a significant difference before and after the launch of the BRI policy. This insignificance could be because some FDI activities (e. g., efficiency-seeking FDI) are more likely to relocate to BRI regions, while others are not experiencing significant shifts. This could result in an insignificant change overall. The study further analyzes different FDI motives to investigate how different investment activities respond to the launch of the BRI.

The results for different FDI motives show that efficiency-seeking FDI, market-seeking FDI and strategic asset-seeking FDI from eastern provinces respond positively to the launch of the BRI as expected. However, trade-supportive investments from the east do not indicate a significant change after the BRI launch. This implies that BRI regions provide opportunities for firms from eastern provinces to geographically diversify their manufacturing and R&D activities. In addition, Chinese firms have leveraged the BRI policy to seek alternative foreign locations in their international market expansion. The more likely Chinese firms originate from the east and have relied heavily on developed nations’ FDI, the more likely these firms will want to broaden their location choice set for further expansions. What is worth mentioning is that R&D activities or strategic asset-seeking FDI do not have to be cutting-edge or ground-breaking discoveries; they can be knowledge applications such as localizing existing products to better fit the host country’s institutional environment or consumer tastes.

Potential reasons for the unsupported hypotheses related to FDI motives may be as follows. Regarding trade-supportive investments, the BRI policy does boost trading activities in BRI regions. Nevertheless, the increase in trading activities comes from western provinces in China rather than from the east. Therefore, the resource dependence effect, namely that firms from eastern provinces may use the BRI to break the over-dependence on Western European and North American firms, may be over-compensated by trade from western provinces.