The Role of Top Management Team in Oversea Location Choice: Evidence from Chinese Firms’ Investments in European Industrial Clusters

ABSTRACT

This article investigates the role of firm’s top management team (TMT) in its location strategy in oversea investment decision. From the perspective of the upper-echelon theory and knowledge-based view, we study how knowledge-related characteristics of TMTs, such as education background, foreign experience and R&D experience affect the firm’s oversea location choice to invest in an industrial cluster. Using data of OFDI cases from Chinese firms to European sub-national regions from 2006 to 2016, we find that: (1) TMT’s education background has a positive effect on firm’s OFDI decision in industrial clusters; (2) TMT’s foreign experience has a positive effect on firm’s OFDI decision in industrial clusters; (3) TMT’s R&D background increases the firm’s likelihood to invest in industrial clusters. Our findings provide implications related to the effects of TMT characteristics on emerging market firms’ oversea investment activities.

KEYWORDS:

• Industrial cluster
• knowledge
• top management team
• emerging market
• oversea location choice

JEL:

• F23
• O32
• G34

1. Introduction

The past 20 years have seen a substantial amount of outward foreign direct investment (OFDI) coming from emerging market multinational firms (EMNEs), which aim to obtain market-related assets, strategic assets and technology resources to increase their innovation capabilities and competitiveness both domestically and overseas (Li, Huang, and Dong 2019). Among all the emerging markets, China has been particularly active in OFDI and become the third largest overseas investor since 2012 (Li and Rengifo 2018). Many Chinese firms are active seekers of strategic assets in developed countries, and more than 40% of China’s OFDI in developed countries flows into Europe (Dreger, Schüler-Zhou, and Schüller 2017).

The outburst of OFDI from emerging markets marks an important change in the global economic pattern and inspires many studies on determinants and performance of emerging market OFDI (Chen et al. 2015). However, most previous literature only considers country-level factors including economic development, political relationship and institutional environment. These studies do not take into account the different geographical, institutional and technological conditions among sub-national regions. In addition, with the rise of the studies on new economic geography, sub-national regions turn out to have more important effects on firm activities than the whole nation, as the differences among regions can be exaggeratedly large even in the same nation. The resources a firm relies on and needs to develop are mainly concentrated in its surrounding environment but not the national environment (Chen et al. 2015). Therefore, it is important to study firm’s location choice in sub-national regions instead of merely conducting the country-level analyses.

Little research studies firm location choices of OFDI at the regional level, not only because the data on regional characteristics of target markets are lacked, but also because it is difficult to divide location choices into two or three distinct options. The concept of industrial cluster gives a good perspective, especially when the firm’s OFDI is to obtain strategic assets, such as technology and knowledge resources. An industrial cluster often refers to agglomeration of firms from the same or related industries in one region (Dreger et al. 2017). Several studies have confirmed the advantages of industrial clusters over other types of locations in facilitating business activities and innovation (Lai et al. 2014). Locating in industrial cluster has many benefits and challenges, which could have different effects on different firms depending on how firms treasure the advantages and how they can deal with the challenges. Li and Bathelt (2018) have argued that firms’ specific characteristics would affect their choice to invest in foreign industrial clusters. Firms which originate from industrial clusters and developed countries are more inclined to choose industrial clusters, as they are born in highly innovative and competitive environments.

To broaden understanding of firm-level factors on firm’s location choices of OFDI, we propose that firms’ location choices in industrial clusters are highly related to the top management team (TMT), who are the major decision makers of firms. We are motivated to assess the role of the TMT and its effect on firm’s oversea location decision in industrial cluster. In this paper, we argue that as the location choice of a firm’s OFDI in industrial clusters is one of the firm’s most important strategic decisions about internationalization and innovation, it should be influenced by TMTs’ characteristics, especially those related to knowledge and risk preferences of the top managers. We propose that TMTs’ knowledge-related characteristics, such as education background, foreign experience and R&D experience have impact on firm’s oversea location strategies in industrial clusters, and the results are supported by the data on Chinese firm-level OFDI to European industrial clusters for the period 2006–2016.

The theoretical contribution of this article rests on the following aspects. First, it contributes to the international business literature by providing more detailed analysis on emerging market firms’ ODFI location choices in industrial clusters. Second, we add to the literature on corporate governance by examining how TMTs’ knowledge-related characteristics affect firms’ willingness to invest in oversea industrial cluster. Our research findings provide implications to deepen the understanding of the roles of TMTs in the establishment of oversea investment location decisions for firms in emerging countries, especially we highlight the role of TMT members with knowledge-related experience in directing firm’s OFDI activities.

2. Theory and Hypothesis Development

2.1. Industrial Clusters and Location Choice

In the age of knowledge economy, the ability to innovate becomes an important advantage and reflects a firm’s ability to manage, maintain, and create knowledge (Grant 1996). The knowledge-based view emphasizes the role of firms as institutions for integrating and generating knowledge. The industrial cluster is a platform for technological innovation, where firms can be involved in innovation activities of the surrounding firms and gain knowledge resources to increase their competitive advantages (Zhong and Tang 2018). Industrial clusters provide the best environment for knowledge creation and application because firms located in the clusters can be exposed to abundant amounts of knowledge sources generated by neighboring firms. The proximity in space makes it easy for firms to gain knowledge faster and cheaper and eventually increase their innovation strength and the long-term competitiveness (Dreger, Schüler-Zhou, and Schüller 2017). As the technology innovation has been speeding up over the past decade, it is difficult for firms to obtain technology competitiveness unless they locate with innovative firms in industrial clusters (Lai et al. 2014).

In spite of the many advantages of industrial clusters, the evidence shows that clusters can get overcrowded when there are too many firms with homogeneous abilities and technologies in the same area. The fierce competition in the region can result in failure of the firms when a few players take most advantages in the industry (Alcácer, Dezső, and Zhao 2015). In the context of foreign investments, the fear of being left behind is even stronger as firms from other countries also face liability of foreignness and would meet more difficulties in absorbing knowledge spillovers when the local firms have already gained enough advanced knowledge (Zhou and Guillen 2015).

Therefore, locating in industrial clusters has both benefits and challenges, which makes locating in industrial clusters a complicated decision. On the one hand, industrial clusters can help the firm’s subsidiaries to gain more information and develop their technologies and knowledge resources faster. On the other hand, the agglomeration of firms and industry-related resources in clusters also increase the intensity of competition. The benefits from locating in clusters can thus be reduced by uncertainty of knowledge management in subsidiaries and may even result in failure of foreign investment and exits from desired locations. When firms make OFDI location choices in industrial clusters, they make trade-offs between the preference for innovation and knowledge exchanges and the pressure of surviving in highly competitive environments.

The development of Chinese industrial clusters is promoted by the economic opening and rapid industrialization in China since the 1970s, which gave births to numerous cities specializing in different manufacturing industries (Fleisher et al. 2010). Compared with foreign enterprise clusters, in the process of development in China’s clusters, the lack of institutional systems makes enterprises rely more on relational resources and external networks to reduce the uncertainties of the environment (Tan and Peng 2003). Enterprises share resources and benefit from industrial clusters through technology spillovers (Guo and Guo 2011). In the process of internationalization and OFDI, Chinese firms also face the liability of foreignness due to the institutional distances between China and host countries, which in turn shapes the unique patterns in location decisions by Chinese firms (Kang and Li 2018). Chinese firms tend to follow their peers instead of getting into fierce competition in clustered region with European firms (De Beule, Somers, and Zhang 2018). Jindra, Hassan, and Cantner (2016) suggest that Chinese firms are attracted by industrial clusters in host markets, which can provide knowledge spillovers and shorten the learning process of Chinese firms.

2.2. TMT and Location Choices

TMT plays particularly important roles in strategic decision-making and resource building for the organizations (Li 2018). Evidence shows that the composition of the TMT can have effects on international diversification of firms (Tihanyi et al. 2000). The TMT’s personal characteristics and preferences can exert profound influences on the investment and location choice of firms through three channels: knowledge, capability and risk preference.

The knowledge stocks and learning abilities of TMT can have impact on location choices in clusters, as the location choices reflect the fit between the TMT characteristics and the level of knowledge novelty they seek to possess. Firms are found to be more likely to invest overseas if they acquire skilled employees with profound knowledge on oversea markets (Asakawa et al. 2018). For TMTs with rich knowledge and knowledge-related skills, industrial clusters provide them with the platform to make information exchange with incumbent firms and steer toward innovation and breakthrough (Casanueva, Castro, and Galán 2013).

The capability of TMT could increase their knowledge on the most suitable choice of location when investing overseas (Asakawa et al. 2018). Specific experience in certain geographical areas and professional domains can help to cultivate these capabilities and reduce the institutional distances between TMT and the host country. TMTs that acquire more returnees and managers with previous experience-based capabilities are more active in launching IPO overseas (Cumming et al. 2016). The capabilities of TMT that are built with unique experience of interacting with unfamiliar and risky areas can lead to bolder location choices satisfying the long-term goals.

The risk preferences of TMT can also have impacts on foreign location choices. Aggressive TMTs are less hesitated to enter unfamiliar markets, as they treat the investment choice as source of opportunity and profits (Jindra, Hassan, and Cantner 2016). The risk preference of top managers can be changed when they obtain more comprehensive information on investment destinations. TMTs with little knowledge stock and capability may have difficulties handling high competition in industrial clusters and the firms are consequently less likely to take risk to survive in the competitive and crowded environments of industrial clusters. Knowledge stock and capability therefore interact with the risk preferences of TMT and lead them to choose the location in industrial clusters.

Based on the arguments above, we find three characteristics of TMT that are closely related to their knowledge, capability and risk preferences, and we propose that these three characteristics of TMT have impacts on firm location choices in industrial clusters.

2.3. TMT Education Background

TMT’s education background is the most explicit feature representing the knowledge, capability and resources a firm can get access to. Receiving education and earning degrees take efforts and time, so it can stand for an effective signal that the top managers have enough knowledge stock, enough endurance and enough wisdom to cope with the challenges a firm will be faced with. Education background has fundamental effects on the cognitive abilities and attitudes of TMT toward innovation because innovation is basically based on new combination of previously accumulated knowledge (Wincent, Thorgren, and Anokhin 2016). In the internationalization process, TMTs with more education background are more able to solve complex and ambiguous problems and are more open to innovative resources.

Receiving more education is related to long-term orientation because few benefits can be seen in the short term (Figlio et al. 2019). Therefore, they are more likely to look for locations that can provide the resources to strengthen long-term competitiveness of the firms. Industrial clusters have more resources and provide more suitable environment for firms to gain accumulative knowledge and technology required for innovation. Higher level of education drives people to explore unlimited amount of information for specific purposes. The higher the degree, the more knowledge one must get exposed to and the deeper he must explore to solve more frontier problems. To operate highly specialized firms and make them strong enough to cooperate and compete with other firms located in an industrial cluster, TMT with higher degrees in education are needed so that they can have coherent communication and collaborate in a highly efficient way during the innovation process. In this sense, TMTs with more education background are more likely to invest in cluster regions.

In reality, many industrial clusters are surrounded by universities and research centers. A region can fully cultivate and utilize the informal communications between higher institutions and firms to provide more academic knowledge for professionals. Knowledge from university is important for firms in many industries to reach breakthrough innovations, increase productivity, and generate product and process innovations (Santamaria and Surroca 2011). Firms that rely on technology place great emphasis on the links between universities (Bellucci and Pennacchio 2016). Consequently, TMTs with more education background would have more advantages at building relationships with universities in industrial clusters and obtain knowledge-related resources beneficial for their firms. Therefore, we propose that:

H1: The top management teams with higher average degrees of education are more likely to locate their oversea investments in industrial clusters.

2.4. TMT Foreign Experience

Foreign experience of the TMT plays an important role in firm internationalization process, because TMT that have set their feet overseas are typically more aware of the innovation activities going on in the world and are more confident that they can understand the core functions of innovation in other countries and turn them into practice (Carpenter and Fredrickson 2001). Managers with foreign experience would be more able to find opportunities and make decisions to invest in significant resources in innovation activities. They would be more attracted by industrial clusters when making location choices. In addition, top managers with foreign experience have broader views, are readily open to new ideas, and are more able to become acclimatized to changes and risks (Yuan and Wen 2018). TMTs with foreign experience are more equipped with global mind-set and may consider their firms as more qualified than their counterparts to catch up with the local business leaders (Carpenter and Fredrickson 2001). Finally, familiar with foreign institutions and markets, TMTs with more foreign experience may prefer to directly join their competitors in the industrial clusters regardless of the fierce competition (Díaz-Fernández, González-Rodríguez, and Simonetti 2015). Thus, we propose that:

H2: The top management teams with more foreign experienced members are more likely to locate their oversea investments in industrial clusters.

2.5. TMT R&D Experience

When entering a new foreign market, foreign firms tend to learn from local firms to obtain R&D resources specific to the location. Industrial clusters provide perfect environments for the foreign firms to obtain such kinds of resources (Qian, Cao, and Takeuchi 2013). Previous working experience in R&D activities brings managers with more sophisticated knowledge specific to the technology area within or related to the firm’s technological domain. Path dependent on the skills and accumulated knowledge mastered before, they are naturally inclined to choose industrial clusters as destinations of OFDI, as the overall knowledge stock of the related knowledge and innovation-related resources in industrial clusters are richer than other regions. It is also easier for TMT with R&D experience to recognize firms as well as places that have strong knowledge resources and technology-related assets. In addition, the experience obtained from R&D activities equip TMTs with more abilities to handle the competition and risks of operating in clusters. In the context of OFDI from the emerging markets, TMTs with R&D background can guide the innovation activities of foreign subsidiaries even if they are far away from the host regions, as the abilities of coping with R&D issues are seldom affected by geographical distance and institutional distance (Asakawa et al. 2018). R&D background also has impacts on TMTs’ risk preference so that they are more persistent in conducting risky activities (Lee et al. 2017). TMTs with R&D background have high receptivity to competition in industrial clusters. They better understand the fierce competition in the technology field and try everything they can to be the first inventor. Moreover, they are more likely to take part in risky and uncertain innovation-related activities. The arguments above suggest that TMTs with R&D backgrounds are more aware of the technology resources located in industrial clusters, are more able to manage technology-related subsidiaries from long distances, and are more able to handle risky and competitive environments like industrial clusters. Therefore, we propose that:

H3: The top management teams with more R&D experienced members are more likely to locate their oversea investments in industrial clusters.

3. Research Design

3.1. Data Source and Sample

We test the hypotheses in the context that Chinese multinational firms establish their subsidiaries in the European countries from 2006 to 2016. We collect firm’s overseas location decision in industrial cluster from European Cluster Observatory, which provides the most comprehensive data on cluster indicators of 382 sub-national regions inside Europe. Indicators are only observed in years from 2003 to 2013, but the past values show no clear change of states between non-cluster and cluster for any region, so we can easily extend the status of a region to 2016. Chinese listed firms reveal the new establishments of their overseas subsidiaries in their annual reports every year. As we want to refine the study to regional and even metropolitan levels, we look up on the exact locations from their official websites, and complement the sample by searching on Google. Finally, we manage to fit 27 industries (see more details about industries in Table S3 via the supplementary document) by looking at text descriptions of each industry noted in Cluster Observatory as well as the classification codes of firms compiled by China Stock Market and Accounting Research Database (CSMAR). We collect TMT-level and firm-level data from CSMAR. To ensure complete data for all indicators, we end with 288 subsidiaries located in EU regions.

3.2. Variables and Methodology

The choice between clusters and non-cluster is a binary problem and needs a cutoff point to distinguish the two. Therefore, we use the location quotient (LQ) to stand for the level of agglomeration of a single industry in a typical region. The typical method to calculate location quotients is by using industrial production data. Another definition of location quotient is based on the number of employments in an industry in a certain region. Evidence has shown no significance in measure effects between the two methods, as they both represent the level of specialization of an industry in one region (Li and Bathelt 2018). Therefore, we specify the location quotient as:

(1) $L{Q}_{ij}=\frac{\frac{x_{ij}}{{\sum }_j{x}_{ij}}}{{\sum }_i\frac{x_{ij}}{{\sum }_i{\sum }_j{x}_{ij}}}$(1)

where $x_{ij}$ refers to the employment of the ith industry in the jth region. We classify the region $j$ as an industrial cluster of industry $i$ if the location quotient of industry $i$ in region $j$ is larger than 1. Likewise, we classify the region as a non-cluster of industry $i$ if the location quotient of industry $i$ in this region is equal or smaller than 1. The values of LQ are taken from European cluster observatory dataset and are based on employment data, and we assign the value of 1 to dependent variable “Cluster” if LQ>1 and 0 otherwise.

We study three main independent variables: TMT education background, TMT foreign experience and TMT R&D experience. For TMT Education Background (EDU), we assign values to the highest degrees of top managers in a firm by the time length of their education. Then we calculate the average value of education years as the sum of each top manager’s education years divided by the number of all TMT members. For TMT foreign experience (FOR), we calculate the percentage of top managers who have foreign working or studying experience in the TMT. If one top manager in TMT used to hold a position abroad or study overseas, we claim that the top manager has foreign experience. For TMT R&D experience (RND), we measure the ratio of TMT members with R&D background to all TMT members. If one manager in TMT has former R&D background, we claim that the top manager has R&D experience.

Finally, we include other control variables that influence oversea location decisions in industrial clusters. Following previous literature, we control for corporate governance factors (TMT average age, major shareholder, foreign shareholders, shares held by TMT, independent directors), firm ownership (SOE), firm-specific variables (firm age, international experience, firm size), host country specifics (knowledge intensity, marketization index).

The measures for dependent variable and three main variables are detailed in Table 1 (see more details about variable definition and data sources for all variables in Table S1 in online supplemental document).

Table 1. Summary of main variables

Variable	Measures
Dependent Variable: Cluster (0–1variable)	Location Quotient$L{Q}_{ij}=\frac{x_{ij}/{\sum }_i{x}_{ij}}{{\sum }_i{x}_{ij}/{\sum }_i{\sum }_j{x}_{ij}}$LQ>1for industrial cluster, “Cluster = 1”;LQ≤1 for non-cluster, “Cluster = 0”
Independent Variable:
Education Background of TMT (EDU)	$EDU=\frac{{\sum }^{\mathbf{E}}}{N}$E is the years of education a top manager receives, N is total number of TMT members
Foreign Experience of TMT (FOR)	$FOR=\frac{{\sum }^{\mathbf{F}}}{N}$F is the number of top managers who has foreign working or studying experience, N is total number of TMT members
R&D experience of TMT (RND)	$RND=\frac{{\sum }^{\mathbf{R}}}{N}$R is the number of top managers who has R&D experience, N is total number of TMT members

To test the hypothesis, we design the following probit regression model:

(2) $\mathrm{l}\mathrm{n}\left(\frac{\mathrm{p}}{1-\mathrm{p}}\right)={\beta }_0+{\beta }_1\ast \mathrm{E}\mathrm{D}{\mathrm{U}}_{\mathrm{i},\mathrm{t}}+{\beta }_2\ast \mathrm{F}\mathrm{O}{\mathrm{R}}_{\mathrm{i},\mathrm{t}}+{\beta }_3\ast \mathrm{R}\mathrm{N}{\mathrm{D}}_{\mathrm{i},\mathrm{t}}+\gamma \ast {\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{\mathrm{i},\mathrm{t}}$(2)

where p stands for the probability of Chinese firms locating their investments in industrial clusters, $i$ and $t$ subscript for firm and year, $ED{U}_{i,t}$, $FO{R}_{i,t}$, $RN{D}_{i,t}$ represent TMT education background, TMT foreign experience and TMT R&D experience. $X_{i,t}$ stands for control variables.

4. Empirical Results

4.1. Descriptive Statistics

The most popular destinations of investments are Germany, Italy, United Kingdom, the Netherlands and France. The average LQ in the United Kingdom is the highest among all countries. Countries that are perceived as more developed, including France, Germany, Italy, Switzerland and Denmark have high average LQ, while less developed countries like Hungary and Spain have low LQ compared with their neighbors. Having collected information on firms’ industries, 70% of the subsidiaries belong to industries with high or medium–high knowledge intensity. Over 30% subsidiaries reside in digital products, heavy manufacturing, electric power and automobile, which are the industries that require huge amounts of technology, expertise and R&D activities (see Tables S2 and S3 in online supplemental document).

Prior to testing the proposed hypotheses, the correlations analysis was conducted to examine possible correlations between the variables, finding no serious issues of multicollinearity (see Table S4 in online supplemental document).

4.2. Probit Regression Results

Table 2 displays the probit regression results for our models. In all models, to account for the time effect, year dummies are controlled. Moreover, we use robust probit regression to obtain unbiased standard error under heterogeneity. Both Model 2 and the full Model 5 show a positive and significant effect of the TMT average education

Table 2. Probit regression results (dependent variable: investment in clusters).

	Model1	Model2	Model3	Model4	Model5
TMTAGE	0.020	0.014	0.012	0.023	0.003
	(0.028)	(0.028)	(0.033)	(0.029)	(0.034)
SHCO	0.003	0.002	0.008	0.002	0.007
	(0.007)	(0.007)	(0.008)	(0.007)	(0.008)
FORSH	−0.006	−0.010	−0.009	−0.007	−0.011
	(0.012)	(0.012)	(0.013)	(0.012)	(0.013)
TMTSH	−0.008	−0.011	−0.011*	−0.009	−0.013*
	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)
OUT	1.116	1.186	−0.058	0.587	−0.074
	(1.845)	(1.808)	(2.004)	(1.876)	(2.034)
SOE	−0.404	−0.646*	−0.458	−0.458	−0.715**
	(0.321)	(0.340)	(0.337)	(0.324)	(0.360)
FAGE	−0.170	−0.099	−0.133	−0.100	−0.003
	(0.213)	(0.220)	(0.236)	(0.216)	(0.249)
SIZE	−0.032	−0.094	−0.141	−0.018	−0.159
	(0.111)	(0.117)	(0.116)	(0.114)	(0.123)
IE	−0.030	−0.033	−0.029	−0.036	−0.029
	(0.028)	(0.027)	(0.029)	(0.029)	(0.029)
KI	−0.080	−0.161	−0.104	−0.087	−0.187
	(0.108)	(0.111)	(0.120)	(0.108)	(0.123)
MKT	0.068	0.094	0.059	0.063	0.077
	(0.091)	(0.092)	(0.094)	(0.092)	(0.096)
EDU		0.268***			0.236**
		(0.095)			(0.102)
FOR			1.587**		1.165*
			(0.650)		(0.72)
RND				0.675**	0.506*
				(0.275)	(0.296)
Constant	−0.151	−3.288	2.826	−0.644	−0.743
	(2.872)	(3.309)	(3.013)	(2.966)	(3.586)
Pseudo R2	0.050	0.086	0.080	0.068	0.116
Log Likelihood	−107.82	−103.75	−95.63	−105.82	−91.82

Stand errors are in parentheses; ***p < .01; **p < .05; *p < .1.

level on firms’ decisions to invest in industrial clusters. These results support our Hypothesis 1. We calculate the average marginal effects, one more year in TMT members’ average education increases the likelihood of investing in an industrial

cluster by 7.9%. Model 3 tests the effect of TMT members’ foreign experience on firms’ location choice in cluster and finds a positive and significant relationship. This result is further supported by Model 5, lending further support for our Hypothesis 2. The average marginal effects show that one additional percentage of TMT with foreign experience increases the likelihood that a firm investing in a cluster by 39%. Model 4 and model 5 show a significant and positive effect of TMT R&D related experience. One additional percentage of TMT who has working experience in R&D functions increases the likelih`ood of a firm investing in an industrial cluster by 17%. These results support our Hypothesis 3. To examine how well the model behaves in predicting investments between clusters and non-clusters, we also looked through the percentage correctly classified in the probit model. The overall correction rate is 67.10%, with 83% of the investments in clusters correctly classified.

To further explore firms’ oversea investment decisions in cluster, we differentiated our sample according to firm size and ownership type. In terms of firm ownership types, we divide the full sample into two groups, state-owned enterprises (SOEs) and private firms (non-SOEs). In terms of firm size, we divide firms into two groups: large firms that are larger than average size and small firms that are smaller than average size. We test the effect of TMT education background, TMT foreign experience and TMT R&D experience on firm’s location choices in industrial clusters in separate sub-samples, results are shown in Table 3. TMT education background, foreign experience and R&D experience have significant positive effects for SOEs, but not for non-SOEs. This result may be due to the fact that SOEs have more policy supports from the government and are more courageous when making overseas investment choices. It is also worth noting that when SOEs get older and further strengthen the government control through ownership, their likelihood of investing in clusters will be offset by the government’s need in other aspects other than knowledge seeking and innovation. In terms of firm size, while TMT foreign experience and R&D experience are all significant and positive for small firms, the effect of foreign experience is insignificant for large firms. Small firms can not only utilize their managers with foreign experience to guide them through the risky process of internationalization, but also fully take advantages of the managers with former R&D experience to help them locate suitable places for innovation-related activities.

Table 3. Probit regression for sub-samples by ownership type and size

	By Ownership Type		By Firm Size
	non-SOE	SOE	Small	Big
	cluster	cluster	cluster	cluster
TMTAGE	0.002	0.143	0.009	−0.103
	(0.041)	(0.210)	(0.047)	−0.091
SHCO	0.006	−0.086	−0.001	0.042***
	(0.009)	(0.075)	(0.011)	−0.013
FORSH	−0.013	0.263**	−0.014	−0.138***
	(0.015)	(0.114)	(0.013)	−0.052
TMTSH	−0.014*	−0.637***	−0.015*	−0.009
	(0.007)	(0.232)	(0.008)	−0.018
OUT	0.199	10.157	−1.753	−0.912
	(2.515)	(13.734)	(3.376)	−2.98
FAGE	0.154	−6.659***	0.087	−1.213**
	(0.307)	(2.447)	(0.308)	−0.588
IE	−0.034	0.049	−0.039	0.02
	(0.040)	(0.146)	(0.049)	−0.048
SIZE	−0.192	0.612	−0.383	−0.551**
	(0.164)	(0.527)	(0.323)	−0.265
KI	−0.190	−1.578**	0.017	−0.794**
	(0.143)	(0.748)	(0.160)	−0.363
MKT	0.048	1.050	0.097	−0.096
	(0.118)	(0.658)	(0.138)	−0.181
EDU	0.155	0.733*	0.059	0.032
	(0.121)	(0.425)	(0.126)	−0.221
FOR	1.331	10.364*	1.824**	1.167
	(0.857)	(5.363)	(0.881)	−1.297
RND	0.553	3.746*	1.327**	0.983*
	(0.441)	(2.386)	(0.578)	−0.578
Constant	0.882	−25.632	6.940	21.427***
	(4.764)	(18.579)	(8.377)	−7.634
Pseudo R2	0.109	0.612	0.139	0.340
Log Likelihood	−66.773	−7.635	−51.780	−27.410

Stand errors are in parentheses; ***p < .01; **p < .05; *p < .1.

4.3. Robustness Tests

Additional analysis is conducted to increase the robustness of the research. First, to test for robustness of data period, we run the main regression using data from 2006 to 2013. Second, we test the effects of the main variables with the total years of TMT education, total number of top managers with foreign experience and total number of top managers with R&D experience, while including TMT size as a control variable in the model. Third, we test the effects of TMT education with two alternative measures for TMT education level: (1) the proportion of top managers with master degree and above; (2) the proportion of top managers with PhD degree and above. Fourth, to control for the industry group, we extend the method of Li and Bathelt (2018) and divide all 27 industries into 7 industry groups, and include industry groups fixed effects in all regressions. Finally, we use the logit regression model to test our model. In general, the results are consistent with our main estimation.

5. Conclusion

Despite the importance of industrial clusters as the sources of knowledge spillover and the platforms for information exchange, studies on firm characteristics that influence their location choices between clusters and non-clusters are still rare. Moreover, the roles of TMT characteristics have never been considered in this problem, even though the knowledge stock and experience of TMT can play strong roles in firms’ oversea location choices. Using the data of Chinese listed firms, we provide new conceptual and empirical insights into the firm oversea investment choices in the sub-national clusters. Specifically, we examine the TMT characteristics that are related to knowledge and technology. We argue that these characteristics have impacts on firms’ investment in foreign clusters. We highlight the roles of TMT education background, foreign experience and R&D experience, because they are highly related to innovation and ability to absorb knowledge and technology spillovers, and could help the TMT to buffer risks and competition when operating in clusters. Through empirical analysis in a setting where Chinese firms choose EU industrial cluster regions for investments, our hypotheses get supported. Our findings suggest that the emerging market firms with higher education degrees, more foreign experienced and R&D experienced TMTs, would be more likely to locate in industrial clusters when expanding overseas. We suggest that shareholders elect top managers with higher education background, more foreign experience and R&D experience, because they could better help the firms to restructure knowledge and risk preferences and to locate in oversea clusters which provide firms with knowledge resources and technology for innovation. In addition, the fact that TMT characteristics have weaker effects on location choices of non-SOEs and larger firms may indicate that the knowledge base and experience of TMT members are not efficiently used in these firms compared to SOEs and smaller firms.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [71972073].