Determinants of China’s outward foreign direct investment to sub-Saharan Africa

(1)

Determinants of China’s Outward Foreign Direct

Investment to Sub-Saharan Africa

Name: Daan Wulffraat Student number: 5965780 Date: 21/06/2014

Supervisor: dr. M. Micevska Scharf Faculty of Economics and Business

(2)

2

Abstract

Since the beginning of this century, China’s outward foreign direct investment (OFDI) has increased over ten fault. The Chinese government has recognized the opportunities to expand China’s global power and influence by making investments abroad. Even though China’s OFDI flows to Sub-Saharan Africa have increased over 20 fault in the last decade, empirical scientific literature has paid little attention to this recent development. This thesis researches the determinants of China’s OFDI flows to Sub-Saharan Africa and analyzes data from 2004-2010 using the following three regression models: the pooled OLS regression model, the Fixed Effects regression model and the random effects regression model. The results of this research show that Chinese outward FDI is attracted to Sub-Saharan African countries with a large market size and a large endowment of natural resources. These results indicate that Chinese firms undertaking OFDI in Sub-Saharan Africa are mainly

(3)

3

Motivation

The subject of this thesis captures many of my interests, because I’m interested both in China's rapid development in striving to become the largest economy in the world and in Africa’s recent growth developments. The research question involves aspects of both of these two contemporary

developments.

1. Introduction

China began to open its borders in the late 1970s. At that time, the Chinese government introduced what it called an ‘open up’ policy and began progressively inviting in businesses from other

countries. This opening up to global markets caused an increase in inward FDI flows year by year. In 2001, China took the next step in its ‘open up’ policy by entering the World Trade Organization (WTO). In anticipation of the accession into the WTO, the Chinese government introduced its ‘go global policy’. This policy encourages Chinese enterprises to make foreign investments and is the first major step towards China’s outward foreign direct investments (OFDI) increase. Chinese policymakers started to see the opportunities that investing abroad brings: “Chinese policymakers have widely recognized outward investment as a necessary stage of growth for Chinese companies and as a precondition of their ability to compete in global markets” (Deng, 2007, p72). These

stimulating policies caused the Chinese economy to grow rapidly. China recently became the second largest economy in the world, next to the US. Most recent predictions state that China will overtake the US as the world’s largest economy by 2020 (World Bank, 2012). Even when global FDI flows declined, China's OFDI kept rising. Since 2002, China’s Outward Foreign Direct Investments grew from $37,2 billion to $509 billion in 2012 (graph 1, UNCTAD, 2013). According to the 2010 Statistical

Bulletin of China’s Outward Direct Foreign Investment, China became the fifth largest economy in the

(5)

5 Graph 1

Source: UNCTAD statistics

China's OFDI flows to African countries have also increased substantially in recent years. Although China’s outward FDI to Africa is still a relatively small percentage of China’s total OFDI (see Chart 1, 4,2% of China’s total FDI in 2010), China's FDI in Africa grew from $500 million in 2003 to $13 billion in 2010. The global run on resources, the increased economic growth rates of African countries in the last decade, and the current perspective of Africa as a continent of financial opportunity have led to a rapid increase of FDI flows to Africa in recent years (Ouma, 2012).

Chart 1

Source: Ministry of Commerce China, 2010

China initially invested in Africa to secure a supply of natural resources, because China did not want to rely on other global markets. Besada, Wang and Whalley (2008) stated that in the last decade China’s FDI flows to Africa have reached many different sectors in many different African countries,

(6)

6 which would indicate that China’s FDI flows to Africa are no longer driven only by a demand for natural resources.

This thesis investigates why some African countries attract more FDI inflows from China than others. The research question will be: what are the determinants of China’s OFDI flows to Sub-Saharan

Africa?

In what way will this thesis contribute to the literature? Most studies focus on China's OFDI in general (Wang, Kafouros & Boateng, 2011; Ramasamy, Yueng & Laforet, 2010; Wiig & Kolstad, 2010). Other studies specifically focus on China's OFDI to Africa or Sub-Saharan Africa, but those studies are non-empirical (Ajakaiye & Kaplinsky, 2009; Besada et al., 2008). None of the studies on China's OFDI in Africa are empirically researching the country-specific characteristics related to the amount of Chinese FDI inflows.

In the empirical part of this thesis, the 2010 Statistical Bulletin of China’s Outward Foreign Direct

Investment will be used, which is a highly-detailed dataset containing data on China’s OFDI

investments with detailed information on region and country. This dataset is published by China's Ministry of Commerce and contains OFDI data from 2004 through 2010. In the econometric model, panel data will be used to research the determinants of China’s OFDI to Sub-Saharan Africa. In the model, only data for Sub-Saharan African countries is used, which means Morocco, Algeria, Tunisia, Libya and Egypt are excluded from the model. This is a commonly used method when analyzing data for Africa.

In the next chapter, the general theories on FDI will be described. In Chapter 3, the general FDI determinants will be discussed. Chapter 4 discusses China and the way in which its OFDI might be different from the OFDI by firms in other source countries. China’s capital market imperfections and China’s institutional environment are specifically discussed. Chapter 5 gives an overview of the existing literature on China’s OFDI determinants. Chapter 6 will be the start of the empirical section of thesis, in which the variables that will be used in the model are introduced. In chapter 7, the data will be analysed and also the research methodology will be described. Chapter 8 deals with the regression results and chapter 9, which is the final chapter, deals with the conclusions and limitations.

(7)

7

2. FDI theories

In order to explain the existing FDI theories, this section starts by giving the definitions of two important concepts used in every FDI theory. These concepts are the multinational enterprise and foreign direct investment.

Dunning and Lundan (2008) give the widely-accepted academic definition of a multinational enterprise (MNE), which is as follows:

“A multinational enterprise (MNE) is an enterprise that engages in foreign direct investment (FDI) and owns, or in some way, controls value-added activities in more than one country” (p.3). This means that an enterprise that undertakes FDI automatically becomes an MNE. According to the OECD (2008) Benchmark Definition of Foreign Direct Investment, a Foreign Direct Investment is:

“a category of cross-border investment made by a resident in one economy with the objective of establishing a lasting interest in an enterprise that is resident in an economy other than that of the direct investor”(p.17).

The home country of the MNE that pursues FDI is called the ‘source country’. The country in which this MNE pursues FDI is called the ‘host country’. If an MNE owns at least 10% of the foreign firm that it invests in, then the investment is identified as a Foreign Direct Investment according to both the OECD and the IMF (Dunning & Lundan, 2008).

2.1 Early FDI theories

There is a wide variety of models and approaches trying to explain the determinants of FDI. No single theory is widely accepted as the best, which leads to a large variety of theories, each of which have their corresponding relevance and value. Faeth (2009) gives an overview of the most important theories of FDI. Early empirical studies were not based on a theory of FDI, because such a theory did not yet exist. In these early studies, companies where asked about their reasons for investing abroad. By way of summary, these early studies found that market factors, mainly market size and market growth, where important determinants of FDI. Cost factors such as labour and the

availability of resources were also found to be important determinants of FDI (Faeth, 2009). The first theoretical models on FDI where based on the Heckscher-Ohlin trade model. This model assumes two countries (home and foreign), two production factors (capital and labour) and two goods. The model assumes a perfect competitive world in which the two countries differ in their relative factor endowments, which leads to relative factor price differences between the two countries. This means that the capital-abundant country would export the capital-intensive good to

(8)

8 the labour-abundant country. In the absence of commodity trade, a firm in the capital-abundant country would move capital to the capital-scarce country, because the capital-investing firm can achieve higher returns on its investment in the capital-scarce country. This is how the Heckscher-Ohlin model explains FDI. The amount of FDI undertaken is positively related to the degree in which a country is capital-abundant (Faeth, 2009).

Over time, theories based on the Heckscher-Ohlin model were heavily criticised because of their strong assumption of perfect competition. Two of the main FDI theories assuming imperfect competition are the theories by Hymer (1976) and Dunning (1979).

2.2 Hymer’s ownership advantages

Hymer (1976) was one of the first to criticize the existing theories that strongly relied on the

assumption of perfect competition. Although there is no single generally accepted theory on FDI, all FDI theories criticizing the assumption of perfect competition agreed that there has to be some form of distortion in order for FDI to exist (Denisia, 2010). Hymer (1976) argued that local firms are better informed about the host country’s market and so, foreign firms needed certain firm-specific

advantages to be able to compete with these local firms. These advantages are called ‘ownership

advantages.’ Examples of these advantages are product differentiation, internal economies of scale,

advantages in technology or subsidies. According to Hymer (1976), these advantages are necessary to counterbalance the disadvantages that come with investing in a foreign market (for example informational, cultural or legal disadvantages). A firm would undertake FDI if the benefits of exploiting the ownership advantages of the firm would outweigh the costs of operating in a foreign country (Hymer, 1976).

2.3 Dunning’s OLI theory

Dunning’s (1979) OLI-theory (ownership, location and internalisation-theory) can also be seen as one of the main theories trying to explain FDI by assuming imperfect competition. Dunning’s OLI theory is a mix of several existing theories, including the theory by Hymer (1976). In this theory, FDI will be undertaken if three types of advantages for an MNE are present. Firstly, the MNE should have

ownership advantages, as explained by Hymer (1976). These are advantages over foreign firms

mainly in the form of intangible assets including patents, knowledge and reputation. Secondly,

(9)

9 good character of ownership advantages. So, according to the OLI-Theory, it must be more beneficial to the firm to internalize these ownership advantages instead of selling them to foreign firms. Thirdly, it should be profitable for the firm to use these advantages in a foreign country because of, for example, lower production costs or tax advantages. These are called location advantages (Dunning, 1979; Faeth, 2009). FDI will be undertaken if these three advantages are present.

Dunning (1979, 1993) also gives three primary motivations of investing abroad. These are as follows: - (foreign) market-seeking FDI

- Efficiency-seeking FDI - Resource-seeking FDI

Buckley, Clegg, Cross, Liu, Voss and Zheng (2007) describe these three motives in depth.

Market-seeking FDI will mainly be undertaken by firms that are located within emerging economies in order

to facilitate exports by host-country producers and to stimulate exports to other large or fast-growing economies. Efficiency-seeking FDI is used to achieve a more efficient production by investing abroad, mostly in search of lower labour costs. Resource-seeking FDI is a foreign direct investment in order to get access to or secure the supply of raw materials and energy sources, especially in the case of emerging economies when there are supply shortages. (Buckley et al., 2007).

2.4 Summary

FDI theories have evolved from the heavily criticised models relying on the strong assumption of perfect competition to models that assume imperfect competition. These models assuming

imperfect competition give a more realistic insight into the determinants of FDI, because they can be linked to the perceived market power of MNE’s.

Faeth (2009) concludes that FDI should not be explained by a single theory but is best explained by a combination of different theories. The empirical evidence shows that each of the different models has a relative value, and so, no theoretical model should replace another; rather, all models should be made available for research.

(10)

10

3. Empirical literature on general FDI determinants

Besides the theoretical literature on FDI, there are also many empirical studies on the determinants of FDI. There is a wide variety of variables that can be used as possible determinants of FDI. Giving an overview of all of these possible variables is beyond the scope of this thesis; however, this thesis will provide an overview of some commonly found determinants of FDI. When comparing many of the studies on FDI determinants, one will notice that many of these studies have conflicting conclusions. To a certain extent, these conflicting conclusions can be explained by the wide

differences in perspectives, methodologies and sample selections being used by the different studies (Chakrabarti, 2001).

3.1 Market variables as determinants

Many studies use market size and market potential variables as FDI determinants. A host country with a large market size brings more opportunities for an MNE to generate increased economies of scale (Buckley et al., 2007). Chakrabarti (2001) lists the many studies that use market variables as potential FDI determinants. He shows that in all of these studies, market size is found to be a significant determinant. “Market size has, by far, been the single most widely accepted as a

significant determinants of FDI flows” (Chakrabarti, 2001, p.96). Most studies that included a market size variable use the host country’s per-capita GDP as a proxy for the country’s market size. Some other studies use absolute GDP as a market size proxy. According to Chakrabarti (2001), per-capita GDP is the commonly used market size indicator. On the other hand, both indicators have their relevance since per-capita GDP and absolute GDP both measure different aspects of the host country’s market size. Per-capita GDP is a proxy for the host country’s income level, while absolute GDP is a proxy for the host country’s economy size (Chakrabarti, 2001). The third commonly used market variable is GDP growth. An economy with a high GDP growth is expected to provide more profitable investment opportunities compared to economies with low GDP growth rates. Although Chakrabarti (2001) shows that most studies find that FDI flows have a positive effect on GDP growth. Some studies also find that FDI flows have an insignificant effect of GDP growth.

Besides these commonly used market indicators, some other studies use alternative market

indicators. Kok & Ersoy (2009) use the annual percentage of GDP per capita growth as a market size variable. They find that FDI flows have a significantly positive effect on the host country’s GDP per capita growth. Carr, Markusen and Maskus (2001) give a different insight by looking at GDP

(11)

11 indicators from both the host country and the source country. Their results show that investments from a source country to a host country increase each country’s GDP. This indicates that both the source country’s and host country’s GDP are determinants of FDI. According to Kok & Ersoy’s findings, the similarity in size of both source and host country is a significant determinant of FDI. They use the variable GDP difference to show that similarity in economic size is a significant determinant of FDI.

3.2 Risk indicators as determinants

Many studies on the determinants of FDI use risk indicators in their empirical model. The risk indicators used in these models are often related to the host country’s political environment. A host country with a politically unstable environment is seen as a risk to the MNE that is trying to pursue FDI in that particular host country. Besides the host country’s political environment, the quality of bureaucratic institutions in a host country is often used as a risk variable. Chueng and Qian (2009) capture different risk components in a risk variable. These components include the host country’s bureaucratic quality, corruption level and level of law and order. They find this risk variable to be a significant determinant for China’s OFDI flows in general. Busse and Hefeker (2005) also relate risk to political indicators. They find that political risk and the host country’s quality of institutions are closely related to the amount of FDI inflows.

For most studies on the determinants of FDI, these risk indicators are not the main focus of the study. A study by Wei (2000) differs from most studies in the sense that it focuses on specific risk indicators. Wei (2000) studies the effect of corruption in a host country on the country’s ability to attract FDI. He uses three different corruption indicators that are based on surveys. He finds that an increase in the perceived corruption level of the host country’s government reduces the inflow of foreign direct investment to that particular country.

3.3 Exchange rate and inflation rate as determinants

The host country’s exchange rate is often used as a possible determinant of FDI (Kok & Ersoy, 2009; Chakrabarti, 2001). As the source country’s exchange rate appreciates, the assets in foreign

countries become relatively cheaper. This would provide more profitable opportunities for investing abroad and would indicate a positive relationship between the exchange rate and FDI (host

(12)

12 Kok and Ersoy (2009) and Chakrabarti (2001) of the studies that use the exchange rate as a possible FDI determinant show that often a negative relationship between the exchange rate and FDI is found. According to Chakrabarti (2001), this can be explained by the currency area hypothesis which holds that if a host country’s currency is weaker, it is less likely for an MNE to invest in that country. This is due to the fact that countries with weak currencies are assumed to have higher exchange rate risk, meaning that the net present value from an investment made in a country with a weak currency is lower due to the increased exchange rate risk. This would then cause the exchange rate to have a negative impact on the amount of FDI inflows into the host country.

Besides the exchange rate, a high and volatile inflation rate can also lead to uncertainty and

increased risk. A high increase in inflation rates leads to a lower valuation of the profits obtained by the MNE (expressed in the local currency). The volatility of the inflation rate also brings uncertainty. A volatile inflation rate can be viewed as an increased risk. In line with this theory, several studies (Li & Liu, 2004; Bajo-Rubio & Sosvilla-Rivero, 1994) find a host country’s unstable or rising inflation rate to be a significant deterrent to FDI inflows.

3.4 Policy variables as determinants

Different policy variables can be included as possible FDI determinants. A commonly used

determinant is the host country’s tax rate. A foreign country’s tax rate will influence future MNE’s profits in its FDI. So, the tax rate influences the present value of a foreign investment. Faeth (2009) divides the possible policy variables that can be used as FDI determinants into the following three types of incentives: fiscal incentives (taxes), financial incentives (for example grants, subsidies or insurances given by the host country’s government) and other incentives (for example a highly subsidized infrastructure by the host countries government).

Besides looking at the effects of corruption, Wei (2000) also includes tax variables to evaluate the effects of taxation on a host country’s ability to attract FDI. In line with his expectations, he finds that an increase in the host country’s tax rate on MNEs reduces the inflows of foreign direct investment into the host country. Faeth (2009) concludes:

“In summary, policy variables such as corporate tax rates, tax concessions, tariffs and other fiscal and financial investment incentives had a significant effect on FDI in a number of studies and should thus be considered as potentially important determinants of FDI.”(p.187).

A concluding remark on the importance of tax variables as determinants of FDI is that in general the effect of tax policy on FDI is significant but small compared to the effects of other determinants (Faeth, 2009).

(13)

13

3.5 Resource variables as determinants

According to Dunning (1979, 2008), one of the main motivations of investing abroad is the resource-seeking motivation. The theory behind the resource-resource-seeking motivation is that firms invest abroad to get access to natural resources. Especially in the case of emerging markets that experience rapid growth rates, which is also the case for China, there is an increasing scarcity of natural resources in the domestic supply. Firms in emerging markets then try to obtain these scarce resources by

investing abroad (Buckley et al., 2007). Although Dunning’s framework is a widely-recognized model, there are not many studies researching the specific influence of a host country’s resource

endowments on FDI in general.

4. China’s OFDI

The question then arises as to whether China is different, and if so, in what way. The following section will discuss the determinants of OFDI for Chinese firms. This section will start with discussing two factors causing possible differences between China’s OFDI and the general theories of FDI. These two factors are China’s capital market imperfections and China’s institutional environment. Capital market imperfections will be discussed and a description of China’s outward investment policy will be given. More specifically, China’s ‘go global’ policy introduced in 2001, which is one of the main drivers behind the boost in China’s OFDI, will be explained. A short explanation on China’s political/economic outward investment structure will also be given, which provides useful

background information for the rest of the section.

4.1 Capital market imperfections

Several studies argue that for emerging economies like China, some differences from the general FDI theory arise, since this general theory is mainly focused on outward FDI by developed economies. In emerging economies like China, capital market imperfections exist, which would mean that for Chinese firms, capital is available at a rate below the market rate (Zhang & Daly, 2011; Buckley et al., 2007). According to both Zhang and Daly (2011) and Buckley et al. (2007), three different

imperfections may arise for China. Firstly, firms that are owned by the state may have access to capital at below market rates. Secondly, due to the inefficiency of the banking system in these

(14)

14 emerging economies, banks may provide soft loans to firms wanting to invest abroad. Thirdly, in emerging economies, a higher percentage of family-owned firms exist. These firms may have access to cheaper capital due to capital provided by family members.

4.2 China’s ‘go global’ policy

China’s institutional environment is another factor that can cause China’s OFDI to be different than the general FDI theories. The investment strategy of an MNE is influenced by the home country’s institutional environment. A very supportive government, which provides subsidies, for example, helps MNEs in obtaining the OLI-advantages and so stimulates investing abroad. On the other hand, these kind of stimulating policies also bring disadvantages. These policies are often accompanied by highly bureaucratic procedures in which the policymakers try to influence the investments (Buckley et al., 2007). Furthermore, China’s OFDI flows are highly influenced by China’s political and economic system. The Chinese government plays a crucial role in shaping China’s OFDI flows. This is due to the fact that China’s industries consist of mostly state-owned enterprises. These enterprises are owned by central or provincial governments. In fact, almost all of Chinas top 500 firms are state-owned (Deng, 2007). China’s state-owned enterprises that want to invest abroad first have to get approval from the Chinese Ministry of Commerce. Privately owned enterprises also have to get state approval for every single investment made abroad, but this approval must be given by authorities on the provincial level (Luo, Xue & Han, 2010).

Now that the role of the Chinese government in shaping China’s OFDI is explained, a more specific overview on the content of this policy will be given. As explained in the previous section, the Chinese government has a high degree of influence on China’s OFDI flows. This influence makes

understanding Chinese government policy essential for understanding the determinants of China’s OFDI.

China’s policy concerning globalization started in the late 1970s. China initially responded passively to the globalization starting in the end of the 1970s. In 1978, China introduced an open door policy which mainly focused on FDI inflows. At the beginning of the 21st_{century, the Chinese government}

recognized the opportunities to expand China’s global power and influence by making investments abroad. Initially China introduced its ‘go global’ policy in 2001. This policy stimulated outward FDI investments by Chinese enterprises. Initially this policy was mainly focused on getting access to and control over natural resources. Due to China’s rapid growth, the country seriously lacked natural resources (Ramasami et al., 2012). Later on, the policy changed to a broader vision by also

(15)

15 stimulating firms in making non-resource-related investments abroad. The Chinese policies for stimulating investment abroad have a very wide scope; however, they can be divided into the following five main categories:

1. Policies to create incentives for OFDI.

2. Government-made administrative procedures to make investing abroad less complicated. 3. Easing capital controls, and so, allowing companies to have easier access to capital. 4. Providing information to find the best investment opportunities.

5. Policies to reduce political and investment risk. Luo et al. (2010).

In sum, China’s policy has transformed from a regulating FDI regime to a Chinese government that started guiding and stimulating enterprises in pursuing OFDI. “The Chinese government has been creating a supportive environment that particularly stimulates strong Chinese firms to invest abroad for the purpose of becoming globally competitive.” (Deng, 2009, p76). Due to this strong

government support, Chinese MNEs have rapidly become globally competitive.

5. Determinants of China’s OFDI

As previously mentioned, the literature on FDI determinants has a very wide scope which is beyond the scope of this thesis. In the following section, the literature will be narrowed down to the specific determinants of China’s outward FDI.

There have been several studies empirically researching the determinants of China's OFDI flows in general. One of the first empirical researches within this topic comes from Buckley et al. (2007). The variables used in the empirical model by Buckley et al. (2007) are derived from the existing

theoretical literature on FDI determinants. Buckley et al. (2007) use the three motives for pursuing FDI as described by Dunning (1979, 1993). For the seeking motive, they use three market-seeking variables (host country’s market size, host market size per capita and host market growth). For the resource-seeking motive, the host country’s endowment of natural resources is used. The efficiency-seeking motive is left out of consideration, because this motive occurs when investors search for lower labour costs. According to Buckley et al. (2007), Chinese investors already face low labour costs in their home country, which makes this motive unlikely and so the efficiency-seeking motive is not included in their model. Other main variables used in the model are political risk, ownership advantages (represented by the host country’s patent rate), host country’s inflation rate, host country’s exchange rate with China, cultural relation to China (measured by the proportion of

(16)

16 ethnic Chinese in the host country’s population), trade relations (measured by imports from and exports to China), policy liberalization (a time dummy used for the years before and after China’s policy liberalization) and the country’s openness to FDI(represented by the total amount of FDI inflows into a host country).

Buckley et al. (2007) finds that the market-seeking motive (only the host country’s market size variable), the resource-seeking motive, political risk, inflation rate, exchange rate, cultural relation to China, trade relation and policy liberalization are significant. Here it should be mentioned that Chinese FDI flows are significantly positively related to the host country’s political risk. This means that Chinese firms are attracted by political risk, which is the opposite of what would be expected and which is also different from the behaviour of firms in industrialized countries. Buckley et al. (2007) attribute this to the low capital costs that Chinese firms face. These low capital costs are the consequence of China’s capital market imperfections. This would give less of an incentive for these enterprises look closer at the host country’s political risk.

Kollstad and Wiig (2012) also study the determinants of China’s outward FDI empirically. They use a more recent dataset (2003-2006) than the dataset used by Buckley et al. (2007). In their model, the following three main variables are used: the host country’s institutional level, the host country’s natural resource endowment and an interaction term between the resource endowment and the institutional level of the host country. They also add a number of control variables that were found to be important in previous studies. In their main regression, they use OLS estimations by computing the averages of China’s OFDI to the host countries over the sample. They also perform a panel data regression by using pooled OLS, which seems to give the same results as their OLS regression. Thirdly, they use a fixed-effects regression, which gives few significant results. Their main finding is the significance of the interaction effect, which can be interpreted as the weaker the host country’s level of institutions, the more Chinese FDI is attracted to that country’s natural resources. They also find their other two main variables to be significant; however, unlike Buckley et al. (2007), Kollstad and Wiig do not find an unconditional effect of the host country’s level of institutions on China’s OFDI.

Zhang and Daly (2011) also study China’s OFDI-using panel data. They selected data from 23 host countries that have been the top destinations of China’s OFDI during the sample period of 2003-2009. They also use a pooled OLS regression to estimate their results. They find the following variables to be significant determinants of China’s OFDI: China’s export to the host country, the host country’s GDP growth and GDP per capita. Different from the study by Buckley et al. (2007), they find the other independent variables including exchange rate with China, host country’s import level,

(17)

17 host country’s inflation rate and the host country’s natural resource endowment to be insignificant. In their conclusion, Zhang and Daly (2011) mention that there are limitations to their model and their advice for future studies is to add a variable considering the role of the host country’s government in attracting Chinese OFDI.

Ramasamy et al. (2012) and Wang et al. (2011) represent two recent, general studies on Chinese OFDI flows. Both of these studies give some interesting insights from a different perspective. Ramasamy et al. (2012) distinguishes between government-owned firms and private firms. This study researches the determinants of China’s OFDI based on ownership. Their study uses six different hypotheses, the first five of which concern Chinese firms in general, while the sixth hypothesis distinguishes between state-owned and privately-owned Chinese firms. The first two hypotheses, each of which question whether Chinese OFDI is attracted to countries with a large supply of natural resources and/or is attracted to countries that are assumed politically risky, are based on previous studies including Buckley et al. (2007) and Kollstad and Wigg (2009). The third hypothesis is about the interaction effect of Chinese firms being more resource-seeking in high risk countries. A confirmation of the fourth and fifth hypotheses could suggest that Chinese firms in general are seeking technology. The sixth hypothesis differentiates between private- and state-owned firms. Their results confirm the first two hypotheses, showing that China’s OFDI is more attracted to resource-rich countries and to countries that are politically risky, which is in line with the results of Buckley et al. (2007); however, the interaction effect is not confirmed. Hypotheses 4 and 5 are partly confirmed showing that Chinese firms in general are attracted by countries that are rich in technologies. The main conclusions that can be drawn from the sixth hypothesis are that both private- and state-owned firms are also market-seeking (GDP size is a significant determinant). Only state-owned firms seem to have an attraction to relatively richer countries, also the investing pattern of private firms show that these firms are relatively more risk-averse than state-owned firms.

Wang et al. (2011) give an interesting and different insight on the determinants of Chinese OFDI by focussing on Chinese firm-specific characteristics. They make an empirical analysis to see if Chinese firm-specific factors such as R&D, advertising and human resources are related to the ability of these firms to invest in foreign markets. Most relevant to the topic of this thesis is their conclusion that state ownership has a significantly positive impact on OFDI by Chinese firms. Their findings indicate that factors such as state involvement and a government’s industrial policy are highly important determinants of the ability of Chinese firms to invest abroad.

(18)

18

6.Variable selection

The literature review part of the thesis shows that many different variables are used as possible determinants of FDI. Also, the different studies on this subject are not unanimous in their conlusions on the significance and sign of the different variables. The variable selection for the empirical part of the thesis will be based on the existing literature on the determinants of China’s OFDI. These studies represent a good initial framework. Because of the wide range of variables that are used in these studies, the variables used in the model will be based on a combination of the variables that are commonly used in the existing literature. Also, some variables used in the model will be allocated to the resource- and the market-seeking motives as described by Dunning (1979, 1993). The efficiency-seeking motive will be left out of the model, as this motive is not relevant for Chinese MNEs. Buckley et al. (2007) mentions that due to the fact that Chinese MNEs are already facing low labour costs in their home country, the effiency-seeking motive is unlikely to determine Chinese OFDI.

The first independent variables that will be included in the model can be linked to the market-seeking motive. Chakrabarti (2001) shows that market size is a widely accepted significant determinant of FDI. The larger the host country’s market size, the better the opportunities for economies of scale. The overview of the studies on FDI determinants by Chakrabarti (2001) shows that most studies use per-capita GDP. According to Chakrabarti (2001), per-capita GDP is a better market size proxy compared to using absolute GDP, because absolute GDP would reflect the population size instead of the market potential of a host country. This is especially true when the host country is a developing country. On the other hand, there is a controversy over which variable should be used as a market-size proxy. Absolute GDP and GDP per capita are two very different measurments for the host country’s market size, yet each have an applied value. Chakrabarti (2001) mentions that per-capita GDP is a measurement of the host country’s income level, while absolute GDP is a proxy for the size of the whole host country’s economy. In this study, per capita-GDP will be used as the main variable reflecting market size. Although this varible will be used as the main market-size proxy, the absolute GDP variable will not be ruled out. A regression replacing GDP per capita with absolute GDP will be included for completeness, and these differing results will be compared.

The second market variable included in the model is GDP growth. Buckley et al. (2007) mentions the growth hypothesis, which holds that rapidly growing economies provide more profitable

(19)

19 and GDP growth are expected to have a positive impact on the amount of Chinese FDI inflows into the host country. Therefore, both GDP size and GDP growth will be included in the model, since they both can be linked to the market-seeking motive.

Market-seeking motive:

- GDP per capita/absolute GDP - GDP growth

The next two variables that are included in the model can be linked to the resource-seeking motive. The early FDI theories already found the availability of natural resources to be an important

determinant of FDI.(Faeth, 2009). In many studies on China’s OFDI flows in general (Buckley et al., 2007; Kollstad & Wiig, 2012; Ramasamy et al., 2012), the host country’s natural resource

endowment was found to be a significant determinant of China’s OFDI. These different studies use different variables in trying to approximate the host country’s availability of natural resources. Buckley et al. (2007) use the host country’s ratio of ore and metal exports to merchandise exports as a resource-seeking variable. Kollstad and Wiig (2012) use the total of fuel, ores and metal exports as a share of GDP as a proxy for the host country’s natural resource endowment. Ramasamy et al. (2012) use the host country’s total export of ores and minerals as a proxy. Based upon these different variables being used as a proxy for the host country’s availability of natural resources, it can be concluded that there is no consensus on a single “best variable” that can be used for the country’s resource availability. In this study, two different variables will be used as a proxy for a country’s natural resource endowment. First of all, the total natural resources rents as a percentage of GDP will be used. According to the World Trade Organization (WTO), “the resource rent of a natural resource is the total revenue that can be generated from the extraction of the natural resource, less the cost of extracting the resource.” The amount of these resource rents will be different among the various host countries, as these host countries differ in the costs of extracting these natural resources. The World Bank gathers the data on these natural resource rents. The total amount of natural resource rents for each host country in each year will be used. According to the World Bank, this total amount consists of the sum of oil rents, natural gas rents, coal rents (hard and soft), mineral rents and forest rents. The other variable that will be used as a proxy for the host country’s natural resource endowment is the host country’s ores and metals export as a percentage of merchandise exports. All studies on China’s OFDI in general mentioned in this section (Buckley et al., 2007; Kollstad & Wiig, 2012; Ramasamy et al., 2012) in some way include the host country’s ores

(20)

20 and metals export in their model as a proxy of the host country’s natural resource endowment. The two variables that will be included in the model and that can be linked to the resource-seeking motive are total natural resources rents as a percentage of GDP and ores and metals export as a percentage of merchandise exports.

Resource-seeking motive:

- Total natural resources rents (% of GDP)

- Ores and metals exports (% of merchandise exports)

Several studies on China’s OFDI have included variables that are in some way related to the host country’s political environment. The African continent is known for its politically risky countries with a below-average quality of institutions. In several studies on Chinese OFDI, variables reflecting political risk or the quality of the host country’s institutions are included. A variable on political risk is used in the studies by Buckley et al. (2007) and Ramasamy et al. (2012). As mentioned above,

Buckley et al. (2007) finds that Chinese FDI flows are significantly positively related to a host country’s political risk, which is an unexpected result. Ramasamy et al. (2012) confirms this result. Both studies use different data sources for their variables used as a proxy of political risk. In the empirical model of this thesis, a variable on political risk will be used as well. The data that will be used for this variable comes from the World Bank’s Worldwide Governance Indicators dataset and the specific variable used from this dataset is the variable for Political stability and absence of

Violence/Terrorism. This variable is defined by the World Bank (2012) as a variable “capturing the

perceptions of the likelihood that the government will be destabilized or overthrown by

unconstitutional or violent means, including politically-motivated violence and terrorism.” Based on indications that many African countries are politically unstable, and also based on previous findings by Buckley et al. (2007) and Ramasamy et al. (2012), this political risk variable will be included in the model.

Another OFDI determinant closely related to the host country’s political environment is the host country’s institutional environment. Besides the risk of a government being destabilized or overthrown, the quality and effectiveness of the current government can also influence the

investment decision of a foreign investor (Kollstad & Wiig, 2012). Therefore, the second variable that will be included in the model as a proxy of a host country’s political environment is a variable on

government effectiveness. Data on this variable is obtained from the World Bank dataset. According

(21)

21 and the degree to which these public services are independent of political pressure. So, the variables related to the host country’s political environment that will be included in the model are the Political stability variable and the government effectiveness variable.

Political environment:

- Political stability

- Government effectiveness

A number of additional variables will also be included in the model. These are commonly used additional variables that have been found important in studies on China’s OFDI in general (Buckley et al., 2007; Kollstad & Wiig, 2012; Ramasamy et al., 2012; Zhang & Daly, 2011). The additional

variables that will be added to the model are the host country’s inflation rate, the exchange rate and a proxy for the host country’s openness. The variable on inflation rates reflects the host country’s annual consumer price index change. The variable for the exchange rate contains annual information on the average exchange rate of the host country’s local currency against the Chinese RMB. The data on both of these variables is from the World Bank development indicators database. The third additional variable used in the model is the host country’s total inward FDI stock as a percentage of the host country’s GDP, which is used as a proxy for the host country’s openness. Data on this variable is obtained from the UNCTAD statistical website.

Additional variables:

- Exchange rate - Inflation rate - Openness

(22)

22

7. Methodology and the data

Now that the variables to be used in the model have been discussed, this Chapter will deal with the regression methodology and the data. The Chapter starts with an analysis of the data. After the data analysis, the different regression methodologies that will be used will be explained.

7.1 The data

The goal of the empirical part of the thesis is to identify the determinants of China’s OFDI to Africa. The dataset that will be used consists of annual panel data over a period of 7 years. The dependent variable will be China’s OFDI. This dataset is published by China's Ministry of Commerce and contains OFDI data from 2004 through 2010. Data on most of the independent variables comes from the World Bank. Only the data on the host country’s total inward FDI stock, which is a variable used as a proxy for the host country’s openness, comes from the UNCTAD statistical website. An overviewof all variables and the associated data sources is given in Table 1. All of the data on the independent variables is measured over the seven-year period from 2003 through 2009, as China’s OFDI is

expected to react with a one-year lag. The total dataset consists of 39 Sub-Saharan African countries. This means that the countries north of the Saharan, namely, Morocco, Algeria, Tunisia, Libya and Egypt, are excluded from the model. Excluding the area of the continent of Africa that lies north of the Saharan dessert is a commonly used method when studying African countries. Because of the Saharan, the northern part of Africa is mainly separated from the Southern part, which has led to differences in the key characteristics between these two areas. South-Africa is also excluded, because, according to the World Bank, this country is the only non-developing country in the dataset.

(23)

23 Table 2 presents the descriptive statistics of the variables that are included in the model. The

dependent variable is China’s OFDI flow (per year in millions of USD) which has a minimum value of 0,01 and a maximum OFDI flow value of 390,35 million USD during the period 2004-2010. Per-capita GDP varies from 108,02 to 23511,10 USD. GDP growth in percentages varies from a minimum of 17,67% - which indicates a decline in the country’s total GDP - to a maximum GDP growth of 37,99%. The ores and metals variable shows that some countries do not export these particular natural resources (minimum value of 0,00), while in other countries, the ores and metals exports level covers a large percentage of total merchandise exports with a maximum observed value of 85,37%. The government effectiveness variable ranges from -1,774 (indicating a very weak government performance) up to 0,727, which indicates a slightly below-average government performance. The variable scope ranges from -2,5 up to 2,5. Lastly, the inflation variable reflecting the host country’s annual consumer price index change has a wide range, from a minimum value of -8,98 (deflation) up to 24411,03. The data consists of an unbalanced panel data set, which means that there are missing observations in the data set. The incompleteness of the dataset causes a reduction in the number of observations that are used in the regression model. From the dataset, a total of 129 complete observations is used for the regression model. For some of the variables, including the inflation variable and per-capita GDP variable, some outliers are observed. In the next section, some of the variables (China’s OFDI, GDPc, GDPg, and Inflation) that could benefit from logarithmic smoothing are transformed into logarithms.

(24)

24

7.2 Correlation Analysis

Firstly, the data will be checked on multicollinearity. Multicollinearity refers to the situation where the independent variables are strongly correlated. Multicollinearity can be a problem, since it increases the standard errors of the coefficients. This may lead to a coefficient being considered insignificant, while in fact, when controlling for multicollinearity, these coefficients might actually be significant. This leads to multicollinearity, making it difficult to assess the effect of the independent variables on the dependent variable. Table 3 presents the correlation matrix in which the

correlations between all of the variables being used in the model are shown. Based upon the data shown in this table, none of the variables seem to be excessively highly correlated.

Table 3. Correlation matrix

To check whether multicollinearity is present, the Variance Inflation Factor (VIF) test is used. The results of this test are presented in table 4. Specifically, the VIF measures correlations among the independent variables to gauge the extent to which they affect one another. (Craney & Surles, 2002). There is no specific threshold value in deciding when the VIF is too large; however, to determine the presence of multicollinearity, VIF>5 or VIF>10 are two commonly used VIF threshold values. (Craney & Surles 2002). The results on the VIF values for the variables used in the model show that, even when the lowest threshold value of VIF>5 is used, none of the VIF values passes the threshold value. This indicates that there is no multicollinearity present in the model.

(25)

25 Table 4 VIF

7.3 Methodology

+ µ

it

9

Openn

it-1

+ µ

The RE regression model assumes that all individual countries have a different but random intercept. This random intercept is given by a constant component β0 and a random component. The random

component represents a combination of both country-specific deviations from this constant - represented as εi - and time-specific deviations from this component, represented as µit.

(27)

27

8. Results

In this section the regression results will be presented and discussed. All of the results are presented in Table 5. Reading the table from left to right, the table displays data from the Fixed Effects

regression, the Random Effects regression and then the Pooled OLS regression. In the fourth regression, the GDP per capita variable is replaced by the absolute GDP variable. As described in Section 6, both variables can be used as a proxy for the country’s market size. Two tests discussed below will be used to determine which of the different regression models is the most appropriate for this study. After that analysis, the results from the chosen regression model will be evaluated. F-test

The different regression models have different coefficients and different significance levels for the variables. In order to find out which model is the most appropriate, the F-test will be conducted. The F-test can be used to find out whether the Fixed Effects model or the pooled OLS model is

preferable. The null hypothesis states that all of the intercepts (and so, the country-specific effects) are the same. This implies that the pooled OLS regression model is appropriate. The alternative hypothesis holds that at least one of the countries has an intercept different from another, which would result in the Fixed Effects model being preferred over the pooled OLS model. Given the hypothesis that all intercepts are the same - F(28,91) = 4,62 - the F-test shows a p-value of 0,0000. This result shows that the null hypothesis should be rejected. Not all countries have the same

intercept, and so, the Fixed Effects (FE) regression model is preferred over the pooled OLS regression model.

Hausman test

The F-test showed that the Fixed Effects model is preferred over the pooled OLS model. Since the Hausman test can be used to identify whether the Fixed Effects model or the Random Effects model is more appropriate, the Hausman test will now be used to identify the most appropriate model to be used in this study. The Hausman test identifies whether there is a systematic difference between the coefficients of the Fixed and Random effects models. The null hypothesis of the Hausman test holds that the difference in coefficients is not systematic, and so, the random effects model would

apparently

be an appropriate model to use.

However,

the actual results of the Hausman test show

a Chi-square of 33,87, resulting in a p-value of 0,0001. This means that the null hypothesis should be rejected. Rejecting the null hypothesis has a further implication, namely, that a Random Effects regression would produce biased results. Therefore, the results of the Hausman test show that the Fixed Effects regression model is the preferred and most appropriate model to be used in this study.

(28)

28 So, the Fixed Effects regression model will be the main focus when interpreting and discussing the regression results.

Regression results

Table 5. Regression results dependent variable lnChina’s OFDI

The results from the different regressions are presented in Table 5. The results of the F-test and the Hausman test confirm that the Fixed Effects (FE) regression is the most appropriate regression model. So, the focus will be on the Fixed Effects regressions when interpreting the results. When looking at the first Fixed Effects regression, several interesting results can be found. First of all, China’s OFDI seems to be driven by the market-seeking motive. This can be concluded from the highly significant GDP per-capita variable (lnGDPc), which is significant at a 0,1% level. A 1% increase in the host country’s GDP per-capita increases the inflow of Chinese FDI by 2,93%. This result is in line with general studies on China’s OFDI in that they also found the market-seeking motive to be a determinant of China’s OFDI (Kollstad & Wiig, 2012; Ramasamy et al., 2012; Buckley et al., 2007; Zhang & Daly, 2011). This result indicates that the host country’s per-capita GDP is also a significant determinant for China’s OFDI to Sub-Saharan Africa. The GDP growth variable (lnGDPg) is

insignificant. This indicates that Chinese investments abroad does not seem to be driven by the growth hypothesis, which holds that rapidly growing economies provide more profitable

(29)

29 opportunities. This result is in line with the results by Ramasami et al. (2012), but different from the studies by both Buckley et al. (2007) and Zhang and Daly (2011). These latter two studies find a significantly positive relationship between GDP growth and China’s OFDI in general. An important remark on the findings by Buckley et al. (2007) and Zhang and Daly (2011) is that both studies use a pooled OLS model. However, the pooled OLS model in this study shows a significance of the GDP growth variable, but as detailed above, this is not the most appropriate model.

The other variable in the first Fixed Effects regression model that shows a significantly positive coefficient is the host country’s ores and metals exports (OnM) as a percentage of merchandise exports. This variable is significant at a 1% level. The host country’s ores and metals export level is linked to the resource-seeking motive. The results show that an increase in the host country’s percentage of exports in ores and metals, which are both natural resources, results in an increase of China’s FDI flows to this country. The ores and metals export variable is linked to the resource-seeking motive, which indicates that China’s OFDI flows to Sub-Saharan Africa are also driven by the resource-seeking motive. On the other hand, the variable of the host country’s natural resource rents (RSC) seems to be insignificant. As mentioned before, this variable is an indication of the revenue that can be generated from extracting these natural resources. The differences in the significance of the two resource-seeking variables can be explained by China’s intention behind obtaining these natural resources. Instead of using these natural resources to generate revenues (supported by the insignificance of the RSC variable), China uses these resources to fill the gap between China’s domestic supply and the demand of natural resources (supported by the

significance of the OnM variable). This result can be explained by China’s lack in natural resources due to China’s rapid growth in the last decade. Obtaining these natural resources has been one of the main goals in the early stages of China’s ‘go global’ policy (Ramasamy et al., 2012).

The other variables in the regression seem to be insignificant as determinants of China’s OFDI to Sub-Saharan African countries. Both variables on the host county’s political environment, including government effectiveness (Goveff) and political stability (Polstab), are insignificant. Although both Buckley et al. (2007) and Ramasamy et al. (2012) find the unexpected result that China’s OFDI in general seems to be attracted to host countries with a riskier and more unstable political

environment, this study does not support their conclusion. Both Wei (2000) and Busse and Hefeker (2005) find a more expected result showing that the host country’s political risk is negatively associated with the host country’s ability of attracting FDI. The insignificance of the government effectiveness and political stability for Chinese FDI to Sub-Saharan Africa could be explained by

(30)

30 Chinese capital market imperfections. These imperfections mean that capital is available at below market rates, which would give less of an incentive for Chinese firms to ‘screen’ the host country’s political environment. This explanation is also mentioned by Buckley et al. (2007).

The additional variables are also found to be insignificant. There is no indication that China’s OFDI seems to be determined by the exchange rate between China’s currency and the host country’s currency (Exch). This is in line with the findings by Zhang and Daly (2011), but different from the findings by Buckley et al. (2007). The overviews on FDI determinants by both Kok and Ersoy (2009) and Chakrabarti (2001) show the controversy of this variable as a determinant of FDI. The results of this study show that the exchange rate is not a significant determinant for China’s OFDI flows to Sub-Saharan Africa. According to the theory, inflation was expected to negatively impact FDI. As a

determinant for China’s OFDI to Sub-Saharan Africa, inflation (lnInfl) is an insignificant variable. Most of the studies on China’s OFDI in general (Kollstad & Wiig, 2012; Ramasamy et al., 2012; Zhang & Daly, 2011) find inflation to be insignificant as well. The variable used as a proxy for the host country’s openness (Openn), which is expected to positively impact FDI, is also insignificant. This result is in line with the studies by Ramasamy et al. (2012) and Kollstad and Wigg (2012). Only Zhang and Daly (2011) find this variable to have a significant and positive impact on China’s OFDI in

general. The insignificance of the additional variables could indicate that Chinese firms are mainly driven by the resource-seeking motivation. This resource-seeking motivation could possibly

‘overrule’ the additional variables as significant determinants of Chinese OFDI to Sub-Saharan Africa.

For the sake of completeness, the fourth regression - which is also a Fixed Effects regression - is included in table 5. In this regression, the per-capita GDP variable is replaced by the absolute GDP variable as a proxy for the host country’s market size. Chakrabarti (2001) shows that these different proxies for the host country’s market size are used by several studies on the determinants of FDI. Both measure different aspects of the host country’s market size. The Fixed Effects regression, including the absolute GDP variable, shows no difference with the Fixed Effects regression using the GDP per-capita variable. Both variables used as a proxy of the host country’s market size have a significant and positive impact on China’s OFDI. The results in table 5 thus show two main

determinants of China’s foreign direct investment flows to Sub-Saharan Africa: the host country’s market size and the host country’s natural resource endowments.

(31)

31

9. Conclusion and Limitations

In this study, the determinants of Chinese outward FDI flows to Sub-Saharan Africa are examined. Existing literature has studied the determinants of China’s OFDI in general. This study contributes to the literature by researching the determinants of Chinese OFDI flows to the continent of Africa in specific. The empirical part of the thesis consists of a panel data analysis, in which three different regression methods are used. From these three different regression methods, the Fixed Effects regression turns out to be the most appropriate model. When focussing on the Fixed Effects regression, China’s OFDI flows to Sub-Saharan Africa seem to have two main determinants. These determinants are the host country’s market size (measured by per-capita GDP or absolute GDP) and the host country’s natural resources (measured by the host country’s export level of ores and metals). These two determinants can be linked to the market-seeking and resource-seeking motives. Firstly, China seems to invest in Sub-Saharan Africa in order to get access to larger markets.

Secondly, China seems to invest in Sub-Saharan Africa to get access and control over natural

resources, since China’s need for natural resources has rapidly increased due to China’s high growth rates. The significance of both the host country’s market size and the host country’s endowments of natural resources seems to be in line with the existing literature on the determinants of China’s OFDI flows in general (Buckley et al., 2007; Kollstad & Wiig, 2012; Ramasamy et al., 2012). Only Zhang and Daly (2011) find the resource-seeking motive to be insignificant. Although the determinants of China’s OFDI flows to Sub-Saharan African seem to be in line with China’s OFDI flows in general, there are also some differences. In contrast to the studies researching China’s OFDI flows in general, in which other variables were found to be significant, this study concludes that China’s OFDI flows to Sub-Saharan Africa still seem to be mainly driven by the resource-seeking motive and market-seeking motive.

There are certain limitations to this thesis. First and foremost, as mentioned in chapter 7, the dataset is incomplete. This means that not all the data is available for each variable involving every Sub-Saharan African country for every year analysed. This results in a decreased number of

observations that could be used for the regression analysis. This lack of data caused the total number of complete observations to be reduced to 129. Secondly, due to the possibility of

endogeneity, the results should always be interpreted cautiously. China’s OFDI flows to Sub-Saharan Africa are still relatively small, which makes the issue of reversed causality, in which China’s OFDI would also affect the explanatory variables, less likely. The panel data gives more control over omitted variables. And so, the issue of endogeneity should be considered, and the results should be interpreted carefully. Thirdly, China’s outward FDI is a very recent development which is rapidly

(32)

32 evolving. This causes the observed time period to be relatively short (7 years). During this time period, the world experienced the effects of the global financial crisis, which began in 2007. Since this crisis has had global consequences, it is very likely that this crisis affected China’s OFDI flows to Sub-Saharan Africa. It would be interesting for future research to keep analysing China’s OFDI flows to Africa with the most recent data available to see if the determinants of these OFDI flows to Africa are changing over time.

(33)

33

References

Ajakaiye, O., & Kaplinsky, R.(2009). China in Africa: A Relationship in Transition. European Journal of

Development Research, 21, 479-484.

Bajo-Rubio, 0., & Sosvilla-Rivero, S. (1994). An Econometric Analysis of Foreign Direct Investment in Spain. Southern Economic Journal, 61(1), 104-120.

Besada, H., Wang, Y., & Whalley, J. (2008). China’s growing economic activity in Africa. Working

paper.Cambridge: National bureau of economic research.

Broadman, H. G. (2008).China and India Go to Africa: New Deals in the Developing World. Foreign

Affairs, 87(2), 95-109.

Buckley, P. J., Clegg, L. J., Cross, A. R., Liu, X., Voss, H., & Zheng, P. (2007). The determinants of Chinese outward foreign direct investment. Journal of International Business Studies, 38(4), 499-518. Busse, M., & Hefeker, C. (2007). Political Risk, Institutions and Foreign Direct Investment.

European Journal of Political Economy, 23(2), 397–415.

Carr, D. L., Markusen, J. R., Maskus, K. E. (1998). Estimating the Knowledge-Capital Model of the Multinational Enterprise. American Economic Review, 91(3), 693-708.

Chakrabarti, A. (2001). The Determinants of Foreign Direct Investment: Sensitivity Analyses of Cross-Country Regressions. Kyklos, 54(1), 89-114.

China's Ministry of Commerce (2010), Statistical Bulletin of China’s Outward Foreign Direct

Investment,http://english.mofcom.gov.cn/aarticle/statistic/foreigninvestment/201109/2011090774 2320.html.

Chueng, Y.W., Qian, X. (2009). Empirics of China’s Outward Direct Investments. Pacific Economic

Review, 14(3), 312-341.

Craney, A.T., & Surles, J.G. (2002). Model-Dependent Variance Inflation Factor Cutoff Values,

Quality Engineering, 14(3), 391-403.

Deng, P. (2007). Investment for strategic resources and its rationale: The Case of outward FDI from Chinese Companies. Business Horizons, 50(1), 72–81.

Deng, P. (2009). Why do Chinese ﬁrms tend to acquire strategic assets in international expansion?. Journal of World Business, 44, 74-84.

Denisia, V. (2010). Foreign Direct Investment Theories: An Overview of the Main FDI Theories.

European Journal of Interdisciplinary Studies, 2(2), 104-110.

Dunning, J.H. (1979). Explaining changing pattern of international production: in defence of the eclectic theory. Oxford Bulletin of Economics and Statistics 41, 269–296.

Determinants of China’s outward foreign direct investment to sub-Saharan Africa