Dependent variable

Number of foreign investment projects


Lists of overseas investment enterprises (agencies) issued by Ministry of Commerce

The number of Chinese enterprises’ investments in 65 countries along “The Belt and Road “from 2003 to 2014―a counting variable used to measure the levels of Chinese enterprises’ investments in along “The Belt and Road”

independent variable and regulated variable

Institutional differences(cgzcfl)

Worldwide governance indicators of World Bank

We use “worldwide governance indicators” (WGI) to assess legal and institutional environment in host countries. WGI is an index number system that includes political stability government effectiveness, regulatory quality, rule of law, voice and accountability and control of corruption. The evaluation range of these six indicators is −2.5 - 2.5. Higher positive scores represent better government administration and institution quality. Negative scores represent bad institution quality. Based on the method of principal component analysis, this paper obtains scores of institution quality aggregative indicators in countries along “The Belt and Road”. Institution differences are the absolute values obtained by host countries’ institution quality score minus the institution quality score of China.

Overseas Chinese network (llaborc)

This data originate from National Bureau of Statistics

Past literature materials used the percentage of ethnic Chinese in national population as proxy variables in network of overseas Chinese, so as to assess their vitality in host countries. This data originates from specialized database and the Yearbook of Chinese Economy in Ohio University. Considering that it is cross-section data, this paper uses the stock number of labor dispatching to 65 countries along “The Belt and Road” from 2011 to 2014. In light of bidirectional causal relationship existed in OFDI and transnational labor flow, in order to avoid endogeneity, laborers dispatched to host countries need to adjust to local environment, the process of taking effects may be delayed. As a result, we deal with the data in the next delayed phase.

Control variable

GDP of host countries(lngdp)

World Development Indicators of World Bank database

On the basis of host countries’ GDP calculated according to dollar prices in 2015, use them as proxy variables to assess market sizes of host countries. Larger market sizes represent more chances for corporate investments. Take the logarithms of related data.

Wage levels of host countries (lnhgdp)

This data originates from International Monetary Fund

This paper uses national income per capita as the replacement variable of labor costs. National income per capita not only measures wealth levels of a nation, but also reveals its wage levels.

China’s export to host countries(lnex)

State Statistics Bureau website

Take the logarithm of China’s export to host countries. In theory, export can promote outward foreign direct investment. As result, China’s increasingly export to host countries can also promote outward foreign direct investment of Chinese enterprises.

Resources of host countries (re)

World Development Indicators of World Bank database

An important reason for Chinese enterprises’ foreign investment is to seek for resources. This variable uses percentages of minerals, metals and fuels in GDO as proxy variables.

Technological level of host countries(highex)

World Development Indicators of World Bank database

This variable uses percentages of high-tech products export in GDO as proxy variables.

Geographical distance(lndistance)

CEPII distance database

Use the spherical distance between two countries’ capitals to represent with the unit of kilometers. Take its logarithm.

China’s GDP(lncgdp)

World Development Indicators of World Bank database

Calculate China’s annual GDP based on dollar prices in 2005.

Cultural distance(cd)

Hofstede official websites

Geert・Hofstede, a Holland scholar, proposed four cultural dimensions, including power distance, uncertainty avoidance, individualism/collectivism, male/female, to assess cultural differences and value orientation in different countries. This paper adopts Kought and Singh’s method [14] , as well as the following computational formula:


calculate cultural differences between China and host countries, which represents cultural difference between China and country, the numerical value of cultural dimension in country, the numerical value of cultural dimension in China, as well as the variance of culture dimension indicator.