The location determinants of Chinese foreign direct investment: the
difference between developing- and developed countries
Alfred Venhuizen
1Supervisor: prof. dr. S. Brakman
MSc. International Economics & Business
University of Groningen
2Faculty of Economics and Business
July 2014
---
Abstract
The rise in Chinese outward foreign direct investment draws the question if there is a difference between Chinese investment motives in developed and developing countries. This paper analyses panel data from 15 developed and 20 developing countries on Chinese outward foreign direct investment from 2003-2010 by adopting a fixed effects estimation. Our results show that market-seeking and efficiency-market-seeking investments are motives to invest in developed countries. Developed countries with high-technology industries protect themselves from Chinese OFDI. Market-seeking and resource-seeking investments are motives to invest in developing countries.
---
Keywords: China, Determinants, Outward foreign direct investment
1I would like to thank my supervisor, prof. dr. S. Brakman, for all his valuable comments and suggestions on earlier drafts of this paper.
2 University of Groningen, Faculty of Economics and Business. Student number: s1890395
2 | P a g e 1. Introduction
The last three decades outward foreign direct investment (OFDI) from emerging market economies has undergone significant changes. Foreign direct investment represents an investment in an enterprise made by a foreign firm, involving a long-term relationship and a certain degree of control over the company by the foreign equity owner (Schüller and Turner, 2005). With outward foreign direct investment we mean the flow of capital from the investors’ country to the host country. In recent years, China has become one of the largest FDI suppliers in the world. In this paper, we examine the host-country specific determinants of Chinese OFDI and if this differs between developed and developing countries.
The rise of outward investment from emerging markets has contributed to the growth in FDI globally (Sauvant et al., 2010). In 1980, global FDI outflows totalled 52 billion US dollars. Emerging markets accounted only for 1% in this figure (see table 1.1). By 2012, global FDI outflows were over 1.3 trillion US dollars and emerging markets accounted for 18%. Remarkable is that China accounts almost for one third of this outward investment from emerging market economies. Before the economic crisis of 2008 global OFDI were over 2 trillion. In 2009 it declined with more than 1 trillion compared to 2007. However, Chinese OFDI was still growing.
Table 1.1: FDI outflows, 1980 – 2012 (US$ billions)
Region 1980 1990 2000 2005 2008 2009 2010 2012 World 51.6 241.4 1,092.4 903.8 2,005.3 1,149.8 1,504.9 1,390.9 Developed economies 48.4 229.6 1,090.8 744.4 1,600.7 828.0 1,029.8 909.4 Emerging economies3 0.7 4.4 82.0 76.2 261.8 211.7 244.5 256.5 China4 0.0 0.8 0.9 12.3 55.9 56.5 68.8 84.2
Source: UNCTAD statistics 2014
As follows from table 1.1, China has become the dominant country for outward FDI from the emerging market economies. China’s development path has been widely recognised as being unique, with gradual privatisation and marketization, massive private capital inflows, and extensive exporting. All this has been achieved without political democratisation (Liu et al., 2005). Most studies of FDI related to China, have focused on China as a location for FDI from other countries, rather than as a source of FDI (Kolstad & Wiig, 2012).But now China is becoming an important source of FDI in the
3 | P a g e world. According to the statistics of UNCTAD, China has become the third largest outward FDI supplier globally (2012). Only the United States (328.9 US billions or 23.6% of the total) and Japan (122.6 US billions or 8.8% of the total) are investing more abroad.
One of the reasons that explain the growth of OFDI may relate to the size of China (Zhang and Daly, 2011). China is enormous in economic terms and a small increase in the propensity to invest abroad could lead to a significant change in global FDI. According to the UNCTAD report (2011), the rise in FDI outflows has been driven by various corporate motives and strategies. They name three: The rising demand for oil and gas and minerals to support their economic growth, rising production costs in China, and market expansion.
Emerging market economies invest in many other developing countries. However, they also increasingly invest in developed countries (Sauvant et al., 2010). The phenomenon of OFDI from emerging market economies to developed countries is interesting to economists (Böwer et al., 2009). According to economic theory, capital will flow from capital abundant countries to capital scarce countries. With this new trend, capital flows from the capital scarce country (emerging market economy) to capital abundant countries (developed countries).
Graph 1.1: FDI outflow from China, 2004 – 2010 (US$ billions)
Source: 2010 Statistical Bulletin of China’s Outward Foreign Direct Investment
Graph 1.1 shows the upward trend of OFDI from China to the rest of the world. In 2004, Chinese OFDI totalled 5.5 billion US dollars. 8% (0.4 billion US dollars) of these flows went to developed
0 10 20 30 40 50 60 70 80 2004 2005 2006 2007 2008 2009 2010
4 | P a g e economies5. By 2010, total Chinese OFDI were almost 69 billion US dollars. Now more than 15% (10.7
billion US dollars) of the OFDI flows are invested in developed economies. This growth of investment in developed countries is interesting. Are Chinese firms investing in these countries for the same reason why they invest in developing countries?
We have two interesting facts that occur in China: (1) Chinese OFDI to the rest of the world is rapidly increasing to one of the largest in the world and (2) they invest more and more in developed countries. These points bring interesting questions. There is by now a large econometric literature on the host country determinants of FDI in general. Since FDI in general is dominated by flows from developed countries, it is an open question whether these results generalize to the growing Chinese outward FDI (Kolstad & Wiig, 2012). Also, Chinese investments are often viewed with a mixture of hope and fear. On the one hand, the input of fresh capital is attractive for host countries and developing countries could benefit from technology transfers. However, developed countries fear the loss of key technological abilities (Amighini et al., 2011). Our main research question we want to answer in this paper is: what are the host country determinants of Chinese outward foreign direct investment and is there a difference between the host country determinants of developing and developed countries? We also examine the influence of the economic crisis of 2008 on the host country determinants of Chinese OFDI. In this paper we use panel data of Chinese OFDI of 35 countries from the period 2003-2010. It is released by the Ministry of Commerce People’s Republic of China in 2007 and 2011 and gives for every specific country the value of Chinese OFDI flows.
The structure of the remainder of this paper is as follows: In section 2 we will give an overview of investments of China around the world. Section 3 provides a literature review. Section 4 will explain the data and methodology used, followed by Section 5, which then presents the empirical results. Section 6 will end with a conclusion.
2. Chinese investments around the world
In this section we examine where Chinese firms have invested around the globe in 2010 and the difference with seven years earlier (2003). This is interesting for our country selection. For our sample, we want to include the developing and developed countries where China is most interested in. We also try to predict the reason why they invest in these countries. The book of Dunning & Lundan (2008) provides four motives for outward FDI: Resource-seeking investment, market-seeking investment, efficiency-seeking investment, and strategic asset-seeking investment. For now, we only use these
5 | P a g e terms to predict the motive of investing in a specific country. In section 3, we will explain the four motives in general.
The total stock of outward foreign direct investment from Chinese firms has increased enormously from 2003 till 2010. In 2003, the Chinese OFDI stock was 33 billion and in 2010 it has increased to 317 billion US dollars. When we look at this by region, several things stand out. To mention first, we excluded Hong Kong from the figures. The OFDI stock in Hong Kong is almost 63% of the total of 317 billion US dollars in 2010. Since Hong Kong is a Special Administrative Region in China and an important financial centre, it is logical that a significant amount of capital has flown to Hong Kong. We also excluded Luxembourg, the British Virgin Islands, and the Cayman Islands. These countries are known as tax havens, and thus investments in these countries have no relevance for our study. Later in this paper, we will show that financial centres like Hong Kong have implications for our data.
Figure 2.1: China’s outward FDI stock by region
Source: 2007 & 2011 Statistical Bulletin of China’s Outward Foreign Direct Investment
The region with most Chinese FDI in 2010 is Asia. It accounts for more than 29 billion US dollars. Compared to 2003, the stock in all regions increased dramatically. But relatively, Asia becomes less important for China. In 2003, the stock of Chinese FDI was 45.1% of the total compared to 40.5% in 2010. So, Chinese firms invest less in the region and more overseas. Singapore hosts the largest amount of OFDI stock: 6 billion US dollars. There are a couple of reasons why Singapore attracts foreign investment: It is a large financial centre, its location, skilled workforce, and low tax rates. So an investment by a Chinese firm in Singapore could have different motives. If we take a look at the other Asian countries that have an OFDI stock of more than 0.5 billion, most of the
0 5000 10000 15000 20000 25000 30000 35000 2003 (4.4 billion) 2010 (71.9 billion)
China's outward FDI stock by region
6 | P a g e investments in those countries have a resource seeking motive6. Only the investments in Macau and
Japan are not resource seeking driven. Macau is a Special administration Region in China and thus attracts capital from China (2.2 billion US dollars). Investment in Japan could be seen as strategic asset seeking behaviour. A Japanese news site says the following about recent investments by Chinese companies in Japan7: “All of these investments share the same three investment motifs: technology,
brands and expertise”. Also efficiency-seeking motives could have influence and it is important to notice that distance to China has an important impact on the OFDI stock; the Asian countries are relatively close to China.
Africa has 18.2% of the Chinese OFDI stock (13 billion US dollars) compared to 11.3% in 2003. This implies that Africa is becoming an important region for Chinese firms. The investments in Africa by Chinese companies have a resource-seeking motive. There are many countries with abundant natural resources. Another resource-seeking motive to invest in Africa is the possible food problem in China in the next decades. At the moment, China is self-sufficient but it is expected that within 20-30 years China has to import grains. With investment in the agricultural sector in Africa, China could import food to meet the needs of their own population in the next decades. In the future is it also possible that Chinese firms move production to African countries due to lower labour costs. This has an efficiency-seeking motive.
The third largest region with Chinese OFDI stock is Europe (13.8% or 9.9 billion US dollars). Compared to the 11.2% in 2003, Europe is becoming more important for China. The largest OFDI stock holder in Europe is Russia with more than 2.7 billion US dollars. Chinese companies mostly invest in Russia in the energy-, chemicals-, and mining sectors. So investment in Russia has a resource seeking motive. Other countries with a Chinese OFDI stock of more than 0.5 billion US dollars in Europe are: Germany (1.5 billion), The Netherlands (0.5 billion), Sweden (1.5 billion), the UK (1.4 billion), and Hungary (0.5 billion). The investments in these countries have a market seeking motive or a strategic asset seeking motive. For example, the Chinese telecommunications manufacturing giant, Huawei8,
builds a distribution centre in Hungary (market-seeking motive) and the Chinese car maker Geely takes over Swedish car maker Volvo9 (strategic-asset-seeking motive). Again, the efficiency-seeking motive
could also be important.
6 See list and examples in Appendix C.
7 | P a g e Oceania accounts for 12% of global Chinese OFDI stock (8.6 billion). Australia is by far the largest recipient: 91.4% of Chinese OFDI stock in Oceania is invested in Australia. China is Australia’s biggest trading partner mainly due to China’s demand for natural gas, coal, and iron ore (resource-seeking motive). Australia is also a large market for China (market-(resource-seeking motive) and Australia is highly developed in the field of technology (strategic asset-seeking motive or efficiency-seeking motive). In 2003, the relative share in total FDI stock in Oceania was 10.8%. So, Oceania became more important for Chinese firms.
North America has a Chinese OFDI stock of 7.8 billion US dollars (10.9%) but we have to mention that there are only three countries in this figure. The relative share has decreased with respect to seven years earlier (2003: 12.6%). The reason for this could be that China is now investing in more countries in other regions; in North America there are only three countries to invest in. The US accounts for 4.8 billion and Canada accounts for 2.6 billion US dollars. Bermuda accounts for 0.3 billion but is known as a tax haven. Investment motives in Canada can be seen as the same as in Australia. Canada has abundant natural resources and a large developed market. Investment in the US is slightly different; the US has natural resources but these are primarily used for their own economy. Thus, only market-seeking, efficiency-seeking and strategic asset-seeking motives are the reason to invest in the US.
Latin America has the lowest amount of Chinese OFDI stock (3.4 billion US dollars or 4.7% in 2010) of the OFDI stock globally. The relative share is lower than in 2003 (9.1%). The reason for this could be that Latin American countries have less to offer than the developing countries in Africa with respect to natural resources. When we zoom in to the countries in Latin America, only the OFDI stocks in Brazil and Peru are noteworthy; both are more than 0.5 billion US dollars. Investing in Brazil could have a resource seeking motive; it has abundant natural resources. It also could be a market seeking motive; economists predict that the economy of Brazil becomes one of the largest in the world. Peru’s main exports are copper, gold, zinc, textiles, and fish meals. Thus, a resource motive for Chinese companies seems clear. For countries in Latin America the efficiency-seeking motive could also be applicable.
3. Literature review
3.1 Outward foreign direct investment:
8 | P a g e
outward FDI is provided in the book of Dunning & Lundan (2008). They identify four types of MNE
activity: Natural resource seekers, market seekers, efficiency seekers, and strategic asset or capability seekers. We set out the characteristics of the four types in the following subsections.
The natural resource seekers:
These enterprises are investing abroad to acquire particular and specific resources of a higher quality at a lower real cost than could be obtained in their home country (if, indeed, they are obtainable at all). These firms are seeking for physical resources of one kind or another. The enterprises engaging in FDI with this motive are primary producers and manufacturing enterprises, from both developed and developing countries, which are driven to engage in FDI by motives of cost minimisation and security of supply sources. The resources they seek include mineral fuels, industrial minerals, metals, and agricultural products.
The market seekers:
These are enterprises that invest in a particular country or region to supply goods or services to markets in these or in neighbouring countries. In most cases, previous to the investment, there were exports to this country. They now invest in this particular country because tariffs or trade barriers made it too costly to export to this country or the size of the country now justifies local production. The market-seeking investment may be undertaken to sustain or protect existing markets or to exploit or promote new markets.
The strategic asset seekers:
The next motive we discuss is the strategic asset seeker. These enterprises usually acquire the assets of foreign corporations to promote their long-term strategic objectives. In this way, they could sustain or improve their global competitiveness. Both experienced and first-time investors use this strategy. The motive for strategic asset-seeking investment is less to exploit specific cost or marketing advantages over their competitors and more to augment the acquiring firm’s global portfolio of physical assets and human competences. This motive for foreign investment is increasingly undertaken by MNEs from emerging economies. An example of this investment was the acquisition of IBM’s PC business by the Chinese firm Lenovo in 2005.
The efficiency seekers:
9 | P a g e common governance of geographically dispersed activities. An example of this motive is to change the production stage to a different country due to lower labour costs.
3.2 Determinants of outward foreign direct investment:
Cai (1999) provides in his paper four motives for Chinese FDI: (1) circumventing trade barriers and help to maintain or increase export share in markets of many developed countries (market-seeking motive). (2) Chinese outward FDI is used to acquire a stable supply of resources, primarily in fisheries, forestry, and mining (resource-seeking motive). (3) Outward FDI is used as an effective channel to obtain foreign technology and management skills (strategic asset-seeking motive). (4) Firms in China use outward FDI as an efficient channel to raise capital. They do this by seeking public listings of their shares on stock exchanges through either acquisitions or new listings and by the use of various debt instruments.
Deng (2004) adds two more motives: (5) a number of Chinese companies, particularly large ones, have engaged in overseas investments for the benefit of risk diversification (efficiency-seeking motive). (6) Finally, Chinese firms invest overseas to gain access to networks of existing companies and brands (strategic asset-seeking motive). The UNCTAD 2011 report argues that rising production costs (efficiency-seeking motive) could explain Chinese OFDI.
Looking at the different motives for OFDI, empirical literature of the determinants of Chinese outward foreign direct investment presents the following results:
Market-seeking investments:
10 | P a g e in this particular country because tariffs or trade barriers made it too costly to export to this country. The literature shows that this is the case for developed countries. Thus:
Hypothesis 1: Chinese OFDI is attracted by countries with a large market. Natural resource-seeking investments:
Buckley et al. (2007) find that the natural resource endowment, measured by ore and metal exports to merchandise exports of host country, has a positive impact on OFDI from Chinese firms to those countries. Kolstad & Wiig (2012) find that Chinese OFDI is attracted to countries with poor institutions and large natural resources. They use the same proxy as Buckley et al. (2007). Cheung & Qian (2009) use one different proxy to measure the natural resource-seeking investment motive between developing and developed countries. They use the host-country’s average wage in the manufacturing sector relative to the Chinese one. The other proxy is the host-country’s ratio of raw material exports (including fuels, ores, and metals) to its total merchandise exports. Both proxies have a positive significant impact on Chinese OFDI. In the paper of Amighini et al. (2011), the authors use an industry-level dataset. They divide the firms in three sectors: manufacturing, natural resources, and services. Natural resource-seeking investments, measured by share of fuels on total exports, are only significant in the natural resources industry sector. In these 4 studies natural resource endowment is positively associated with Chinese OFDI, for developing and developed countries. We therefore derive the following hypothesis:
Hypothesis 2: Chinese OFDI is attracted by countries with large natural resources. Strategic asset-seeking investments:
In two papers about Chinese OFDI the strategic asset-seeking motive for investment is examined. The Buckley et al. (2007) paper does not find evidence for this motive. They used annual patent registrations in the host country as a proxy for strategic asset-seeking investment. The Amighini et al. (2011) paper finds evidence of strategic asset-seeking investments in the manufacturing- and service sector. Both proxies ((1) a dummy, 1 if R&D expenditures on GDP more than 1%, and (2) human capital (measured by secondary gross enrolment rate)) have a positive significant impact on Chinese OFDI. Normally, developing countries do not have the strategic assets firms are looking for. Therefore, we derive the following hypothesis:
11 | P a g e Efficiency-seeking investments:
Efficiency-seeking investments were not considered as relevant in the case of China. China has low labour costs and investing abroad for even lower labour costs was not seen as logical. To our knowledge, there are no empirical studies with the efficiency-seeking motive for Chinese outward foreign direct investment. However, the Cheung & Qian (2009) paper uses labour costs as a proxy for natural resources-seeking investments. This could be seen as an efficiency-seeking proxy. So for developing countries, lower labour costs could be a motive for Chinese firms to invest in those countries.
On the other hand, in developed countries the wages are higher. This could imply that they are more efficient. Thus, investing in developed countries could also have an efficiency-seeking motive. We therefore derive the following two hypotheses:
Hypothesis 4a: Chinese OFDI is attracted by developing countries with lower labour costs compared
to China.
Hypothesis 4b: Chinese OFDI is attracted by developed countries with an efficiency advantage over China.
12 | P a g e
Table 3.1: Literature on Chinese OFDI Author(s) (year) Period & data Method Included investment types Findings Determinants Buckley et al. (2007) 1984-2001 Firm-level Panel data (1) Market-, (2) resource-, (3) strategic asset-seeking investment Market- and resource-seeking investment positive impact, strategic asset-seeking investment no impact (1) measured by GDP of host country and by exports from China to host country, (2)
measured by ore and metal exports to merchandise exports of host country, (3) measured by annual patent registrations in host country Cheung & Qian (2009) 1991-2005 Firm-level Panel data (1) Market-, (2) resource -seeking investment Both investment types have an positive impact on Chinese OFDI (1) measured by GDP of host country, (2) measured by host countries average wage relative to Chinese wages and host country’s ratio of raw material exports Kolstad & Wiig (2012) 2003-2006 Country-level Panel data (1)Resource-seeking investment Resource-seeking investment has an positive impact (1) measured by fuels, ores, and metals exports as share of GDP of host country Amighini et al. (2011) 2003-2008 Industry- level Panel data (1) Market-, (2) resource-, (3) strategic asset-seeking investment
All three investment types have an positive impact (in different sectors)
(1) measured by GDP and exports from China to host country, (2)
13 | P a g e 4. Data and Methodology
To examine the different investment motives between developed and developing countries we use a dataset that contains 35 countries. We use 15 developed and 20 developing countries10. We have
selected these countries for two reasons. First, as follows from chapter 2, we have tried to select the countries which received the most FDI inflows from China. Apparently, these countries are the most interesting for Chinese companies to invest in. After this, we dropped some countries because of lacking data. We obtained a strongly balanced panel data set with 280 observations with the time period 2003-2010.
4.1 Model specification & data collection
To analyse the location determinants of Chinese OFDI and the difference in investment motives of Chinese firms between developed en developing countries we derive the following econometric model:
𝑙𝑛𝐶ℎ𝑖𝑛𝑒𝑠𝑒 𝑜𝑢𝑡𝑤𝑎𝑟𝑑 𝐹𝐷𝐼𝑖𝑡
= 𝛼𝑖𝑡 + 𝛽1𝐸𝑋𝑃𝑂𝑅𝑇𝑆𝑖𝑡 + 𝛽2𝑅𝐸𝑆𝑂𝑈𝑅𝐶𝐸𝑆𝑖𝑡+ 𝛽3𝑇𝐸𝐶𝐻𝑁𝑂𝐿𝑂𝐺𝑌𝑖𝑡
+ 𝛽4𝐸𝐹𝐹𝐼𝐶𝐼𝐸𝑁𝐶𝑌𝑖𝑡+ 𝛽5𝑃𝑂𝐿𝐼𝑇𝐼𝐶𝑆𝑖𝑡+ 𝛽6𝐶𝑂𝑅𝑅𝑈𝑃𝑇𝐼𝑂𝑁𝑖𝑡 + 𝛽7𝐷𝐼𝑆𝑇𝑖+ 𝜀𝑖𝑡 We use a Log-Linear function. The function has a logarithmic term on the left-hand side of the equation and has untransformed variables on the right-hand side. This is a common method to study the determinants of foreign direct investment. Similar studies use the same approach (For example, see Anwar et al. (2008), Buckley et al. (2007), Stein & Daude (2001)) Since it has been an effective method to study the determinants of foreign direct investment, we think it is an appropriate method to use in our study. It has a market-seeking component, a resource-seeking component, a strategic asset-seeking component, and an efficiency component. We explain all the variables the following subsections.
Dependent variable:
The dependent variable 𝑙𝑛𝐶ℎ𝑖𝑛𝑒𝑠𝑒 𝑜𝑢𝑡𝑤𝑎𝑟𝑑 𝐹𝐷𝐼𝑖𝑡 is the logarithm of the foreign direct
investment flow in million US dollars from China to host country 𝑖 at time 𝑡. The data is collected from the “Statistical bulletin of China’s outward foreign direct investment” with the time period 2003-2010. This bulletin is issued by the Ministry of Commerce, National Bureau of Statistics, and the
10 Developed countries: The Netherlands, Germany, The United Kingdom, Sweden, France, Italy, Ireland,
Switzerland, Spain, The United States, Canada, Australia, Japan, South Korea, and New Zealand.
14 | P a g e Administration of Foreign Exchange of People’s Republic of China. They only publish data since 2003, and the available data continues until 2010. Similar data is used in the article of Kolstad & Wiig (2012). Independent variables:
Outward foreign direct investment literature identifies the most common motives to invest abroad as market-seeking, resource-seeking, efficiency-seeking, and strategic-asset-seeking. In the following section, we describe the variables which are included in this study.
The market-seeking component:
As we have shown in the literature part of this paper, two proxies for market-seeking investment have been significant explaining Chinese OFDI. In the studies from Buckley et al. (2007), Cheung & Qian (2009), and Amighini et al. (2011), they use the host countries GDP as a proxy for market-seeking investments.
The other proxy that has shown significance is exports from China to the host country. It is used in the Buckley et al. (2007) paper and in the Amighini et al. (2011) paper. It also connects to the theory from Dunning & Lundan (2008). We found that GDP of the host country is highly correlated with exports (more than .90). To avoid multicollinearity issues, we decided to run two regressions; our main function with the variable exports and one with the variable GDP. Variable 𝐸𝑋𝑃𝑂𝑅𝑇𝑆𝑖t tries to explain market-seeking investments. It is the average export value of the three years before the year of investment. So, the export value in 2003 is the average export value from the years 2000, 2001, and 2002 in billion current US dollars from China to host country 𝑖 at time 𝑡 and is collected from the United Nations commodity trade statistics. Variable 𝐺𝐷𝑃𝑖t is the GDP value in billion current US dollars of the host country 𝑖 at time 𝑡 and is collected from the World Bank development Indicators.
The resource-seeking component:
All four studies that examined Chinese OFDI found a significant motive for resource investments of Chinese companies in both developing and developed countries. They all used a similar proxy: fuel or ore and metal exports as share of an export figure. We include the following proxy for the resource-seeking component: 𝑅𝐸𝑆𝑂𝑈𝑅𝐶𝐸𝑆𝑖𝑡. We use the resource endowment rate. That is:
15 | P a g e The strategic asset-seeking component:
Previous studies (Buckley et al. 2007, Amighini et al. 2011) on outward foreign direct investment from China used the proxy’s annual patent registrations in host country or the R&D expenditure by the host country. Since we are examining developing countries, it is hard to get similar data on these countries. Thus, we introduce a new proxy that fits well to the Chinese investment drift: High technology exports by the host country. These exports are products with high R&D intensity, such as in aerospace, computers, pharmaceuticals, scientific instruments, and electrical machinery. Countries with high technology products such as aerospace are interesting for China. For example, China wants to compete with other superpowers in space. So the variable 𝑇𝐸𝐶𝐻𝑁𝑂𝐿𝑂𝐺𝑌𝑖𝑡 is: high
technology exports in billion current US dollars of country 𝑖 at time 𝑡. The data is collected from the World Bank Development Indicators.
The efficiency-seeking component:
The literature on Chinese OFDI does not include an efficiency seeking component to their studies. This is because of the low cost of labour in China. They say that there is no reason for Chinese firms to invest abroad to have the gains of lower labour costs. The UNCTAD (2011) report argues that more and more Chinese firms are investing abroad because of rising labour costs in China. Thus, we want to include a variable to check if Chinese OFDI has an efficiency seeking motive. Finding a proxy that is suitable for our study is difficult because of the lack of data on our 20 developing countries. For example, some studies include as a proxy production costs of the host country. For some countries in our sample this data lacks. We decided to use the difference of GDP per capita between the host country and China as our proxy for efficiency seeking investments. We believe that GDP per capita is closely related to wages in a country. The higher the difference between GDP per capita, the less is China willing to invest in these countries for efficiency seeking reasons. When the number becomes negative (implying that the host country has a lower GDP per capita and thus lower wages), the more China is willing to invest in the host country. 𝐸𝐹𝐹𝐼𝐶𝐼𝐸𝑁𝐶𝑌𝑖𝑡 is the difference of GDP per capita (in
current US dollars) of country 𝑖 and China at time 𝑡. The data is collected from the World Bank Development Indicators.
Control variables:
Political stability in the host country:
16 | P a g e from investing in a country with high political risk because of potential sunk costs (Buckley et al., 2007). Thus, high political risk is generally associated with low FDI inflow. We collected the data from the International Country Risk Guide from the Political Risk Services. 𝑃𝑂𝐿𝐼𝑇𝐼𝐶𝑆𝑖𝑡 is the value of political
risk in of country 𝑖 at time 𝑡. The higher the value, the lower political risk is present. Corruption in the host country:
A major problem in developing countries is corruption. According to Hines (1995), corrupt countries attract less FDI. But for Chinese companies this could also be an opening to gain access to the natural resources of that country. In general, corruption is associated with low FDI inflow. We collected the data from the International Country Risk Guide from the Political Risk Services. 𝐶𝑂𝑅𝑅𝑈𝑃𝑇𝐼𝑂𝑁𝑖𝑡 is the value of political risk in of country 𝑖 at time 𝑡. The higher the value, the less
corruption is present.
Distance between China and host country:
Dunning & Lundan (2008) suggest that firms tend to invest in neighbouring territories or in those with which they have the closest economic, political, linguistic, and cultural ties. For example, in 2000 69% of Canadian investment was elsewhere on the American continent, while 50% of Indian investment was elsewhere in Asia. Thus, we include a control variable that measures distance between the capital city of China (Beijing) and the capital city of the host country. We collected this data from geobytes.com. 𝐷𝐼𝑆𝑇𝑖 is the distance in km between the capital city of country 𝑖 and Beijing.
Data issues
There are some issues with the data. The large investments made in Hong Kong and Luxembourg could first be used to set up a foreign affiliate in that country. Then, that affiliate invests in other European countries. With this investment method, they avoid taxes of the host country. Countries with a higher tax rate could be undervalued in the data.
A second issue are the negative values of the OFDI flow. When we take the logarithm of those variables, these values are dropped. But this is only a minor problem: the number of dropped observations is relatively small11. Secondly, disinvestments may well be driven by other determinants
and thus not important for our study.
17 | P a g e
4.2 Estimation
We use a strongly balanced panel dataset that contains 280 observations for 35 panel groups with an 8 year time period. In appendix D, Tables A, B, and C show the descriptive statistics for the whole sample, the developing countries, and the developed countries, respectively. Tables D, E, and F show the correlation matrices for the whole sample, the developing countries, and the developed countries, respectively.
To test the dataset on multicollinearity, we used the Variance Inflation Factor. When multicollinearity occurs, the least squares estimator is not defined (Hill et al., 2011). With values >5, multicollinearity exists. In this dataset, multicollinearity is not a problem. The highest value of the VIF estimate is 3.07 as shown in table 4.1.
Table 4.1: VIF tests
Whole sample Developing countries Developed countries
Variables VIF 1/VIF VIF 1/VIF VIF 1/VIF
Technology 3.07 0.325 1.53 0.655 2.76 0.362 Efficiency 2.94 0.340 1.32 0.759 1.41 0.710 Corruption 2.43 0.412 1.28 0.784 1.59 0.629 Exports 2.28 0.439 1.64 0.608 2.31 0.433 Resources 1.37 0.732 1.39 0.717 1.39 0.721 Political 1.17 0.853 1.22 0.818 1.40 0.715 Distance 1.16 0.862 1.33 0.753 1.70 0.587 Mean VIF 2.06 1.39 1.79
To analyse panel data on OFDI, three methods are commonly used in the literature: Pooled OLS, Fixed effects, and Random effects. Pooled OLS is one where the data on different individuals are simply pooled together with no provision for individual differences that might lead to different coefficients (Hill et al., 2011). Futures of the fixed effects model are that behavioural differences between individuals are assumed to be captured by the intercept. Individual intercepts are included to “control” for individual-specific, time-invariant characteristics (Hill et al., 2011). The random effects model holds the same assumption, but also recognize that the individuals in our sample were randomly selected, and thus we treat the individual differences as random rather than fixed. We performed a Hausman test to see which model is more efficient for our dataset. The H0 is that the preferred model is random effects. In appendix D the results of the three Hausman tests (whole sample, developing countries sample, and developed countries sample) are included (Tables G, H and I).
18 | P a g e preferred estimation method. In Appendix D, Table J, we included the result of the test. We see that the test is significant. This implies that we reject the H0 and conclude that random effects is preferred. Since fixed effects is preferred above random effects (Hausman test), we use a fixed effects regression method to estimate our model. We do not include a Breusch-Pagan Lagrange multiplier test for the developing- and developed sample. For comparison reasons, we now use the fixed effects method for all three samples.
When we use the fixed effects method, we have to test our samples for a couple of possible problems. First, we test our samples on autocorrelation. Autocorrelation is the correlation of a variable itself over time. This problem biases the standard errors and causes the results to be less efficient (Hill et al., 2011). Here, we use a Wooldridge test for autocorrelation. H0: no first-order autocorrelation. For the whole sample and developing countries the test is significant. This implies that we reject the H0 and conclude there exists autocorrelation. The results of the tests are presented in appendix D, table K.
Secondly, we test for heteroskedasticity. Heteroskedasticity exists when the variances for all observations are not the same (Hill et al., 2011). To test for heteroskedasticity we use the modified Wald test. The H0 is that no heteroskedasticity exists. As we can see in our test results in appendix D, table L, both tests are significant. This implies that we reject the H0 and conclude that heteroskedasticity exists in our samples.
The last test we perform is to test for cross sectional independence. A standard assumption in panel data models is that the error terms are independent across cross-sections. With Pesaran’s test we can test if the error terms are independent across cross-sections. H0: the error terms are independent across cross-sections. For the whole sample and the developed countries sample we reject the H0 because the test is significant (Results: see appendix D, Table M). We can conclude that the error terms are dependent across cross-sections.
19 | P a g e 5. Results
The estimation results of our model are shown in Table 5.1. We used a fixed effects estimation with Driscoll & Kraay (1998) standard errors as presented by Hoechle (2007). The control variable distance is omitted in our fixed effects model because the variable is time invariant.
Table 5.1: Determinants of Chinese foreign direct investment, fixed effects estimation
Variables Whole Developing Developed
Export .01959 *** .13862 *** .01718 *** (.00459) (.02691) (.00420) Resources .16204 *** .17516 *** .03201 (.03132) (.03816) (.03034) Technology -.00488 .09099 * -.01721 *** (.00574) (.04389) (.00348) Efficiency .00007 * .00003 .00014 * (.00004) (.00007) (.00006) Political -.02730 .06326 -.08980 * (.04033) (.05584) (.04327) Corruption .02953 -.00232 .03637 ** (.00959) (.01184) (.01029) Within R² .309 .406 .346 Observations 262 151 111
The column labelled ‘whole’ gives results based on data from both developing and developed countries. The ‘developing’ and ‘developed’ columns, respectively, provide results based on data from developing and developed countries. Driscoll & Kraay (1998) standard errors are in parentheses. ***, **, and * denote significance at the 1, 5, and 10% levels, respectively.
The market-seeking variable export is significant for developing and developed countries. More exports in previous years have a positive influence on Chinese OFDI (hypothesis 1). When there the export number is large to a country, there is a market for these Chinese products. With exporting, trade barriers may cause extra costs for the Chinese firm. So, investing in these countries could circumvent these costs. Chinese firms invest with a market-seeking motive. Previous studies support this finding of Chinese market-seeking investments measured by exports (Buckley et al. (2007), Amighini et al. (2011)).
20 | P a g e explained by investments in developing countries, and not in developed countries. Cheung & Qian (2009) find also a significant effect but, again, not especially for developing countries.
For the strategic asset-seeking motive we used the variable Technology. Surprisingly, for the developing countries it has a significant positive impact on Chinese OFDI. But the coefficient is almost zero. Thus, the level of technology in developing countries is not very important for Chinese FDI. Technology has a negative significant impact on Chinese OFDI in developed countries. Therefore, we have to reject hypothesis 3. An explanation could be that developed countries with a large high-technology industry, protect their domestic industries from Chinese takeovers or Greenfield investments. This is against the findings of Amighini et al. (2011). They find that the R&D expenditures and human capital of a country have a positive impact on Chinese OFDI. The difference with our study is that the findings from Amighini et al. (2011) are found in an industry specific sample. Our OFDI data is the complete flow to a country in a year.
The variable Efficiency stands for the difference between GDP per capita of the host country and the GDP per capita in China for a given year. It measures the efficiency-seeking motive of Chinese firms. The coefficient is positive but not significant for developing countries. The sign is not what we expected; China does not tend to invest more in developing countries with a lower GDP per capita than China. Thus, we reject hypothesis 4a. We cannot conclude that Chinese firms invest particular in these countries for efficiency-seeking motives because the coefficient is not significant. For developed countries the value is positive and significant as we expected (hypothesis 4b). Countries with higher GDP per capita (and thus higher wages), attract Chinese OFDI. Countries with a higher GDP per capita are more efficient than other countries, thus Chinese firms invest in these countries to learn. However, GDP per capita can also be seen as a proxy for market-seeking motives for Chinese OFDI. With the assumed higher wages, more capital is present to spend. This implies a large market for Chinese firms. Previous literature on Chinese OFDI has not taken into account the efficiency-seeking motive of investments. Thus, we cannot compare our results with other studies.
The control variable Politics is only significant for developed countries. For the developing countries, the sign is positive and for developed countries negative. This is reasonable, in developing countries it is more important to have a stable political environment than in developed countries. In developed countries, the chance that your firm is expropriated is lower than in developing countries. An unstable political situation in developed countries could be an opening to investments that are normally not possible for Chinese firms.
21 | P a g e issue. In developed countries, corruption is very bad for the reputation of your business. On the other hand, the sign of the coefficient is negative for developing countries. As mentioned before, corruption could be a way to get access to natural resources of a developing country. Since the coefficient is not significant, we cannot draw this conclusion.
We now perform sensitivity analyses. First, we run the regression with the variable GDP instead of Exports. The literature on Chinese OFDI supports the positive impact of GDP on Chinese OFDI. The results are presented in Table 5.2.
Table 5.2: Determinants of Chinese foreign direct investment, GDP included, fixed effects estimation
Variables Developing Developed
GDP .00182 * .00068 *** (.00082) (.00017) Resources .17787 ** .04411 (.05235) (.03692) Technology .11860 -.02224 *** (.07092) (.00264) Efficiency .00001 .00013 (.00017) (.00007) Political .03654 -.09046 * (.05812) (.04664) Corruption .00977 .03011 ** (.00565) (.01020) Within R² .0369 .326 Observations 151 111
The ‘developing’ and ‘developed’ columns, respectively, provide results based on data from developing and developed countries. Driscoll & Kraay (1998) standard errors are in parentheses. ***, **, and * denote significance at the 1, 5, and 10% levels, respectively.
For both developing and developed countries, GDP has a positive significant impact on Chinese OFDI flows. A higher GDP means a larger market potential and thus investments in these countries are market-seeking investments. This is supported by previous literature on Chinese OFDI (Buckley et al. (2007), Cheung & Qian (2009), Amighini et al. (2011)). However, the coefficients for the variable Exports for developing and developed countries in our main regression are higher than the coefficients of GDP. This implies that exports are more important than GDP for Chinese firms.
In comparison with the regression with the variable Exports included, Technology is no longer significant for developing countries. This seems in line with the expectation; developing countries are not known for their technological abilities. Again, resources of the host country are an important reason to invest in developing countries.
22 | P a g e Chinese firms are investing in developed countries for efficiency-seeking reasons. The other variables are consistent with our main regression.
We now perform a regression with our main variables only; Exports, Resources, Technology, and Efficiency. We take the assumption that only these variables influence Chinese OFDI flows to the host country. The results are presented in Table 5.3.
Table 5.3: Determinants of Chinese foreign direct investment, only main variables, fixed effects estimation
Variables Developing Developed
Exports .11703 *** .01765 *** (.01815) (.00347) Resources .17351 *** .03993 (.03785) (.03656) Technology .07255 * -.00986 (.03173) (.00562) Efficiency .00006 .00014 * (.00010) (.00007) Within R² .390 .298 Observations 151 111
The ‘developing’ and ‘developed’ columns, respectively, provide results based on data from developing and developed countries. Driscoll & Kraay (1998) standard errors are in parentheses. ***, **, and * denote significance at the 1, 5, and 10% levels, respectively.
Without our control variables, the results show the same significant variables for developing countries as in our main regression. Previous exports, natural resources, and technology are motives to invest in developing countries. So, the same arguments hold as for our main regression. For Chinese firms, previous export to developed countries are a motive to invest in those countries. The difference with our main regression is that Technology no longer has a (negative) significant impact on Chinese OFDI. The coefficient is still negative, but without our control variables, we cannot say that Chinese firms invest less in countries with high technologic abilities.
As mentioned in chapter 1, global FDI flows decreased dramatically during the financial crisis of 2008. However, Chinese FDI was still growing. It is interesting to see if the financial crisis changed the investment behaviour of Chinese firms. Therefore, we split the data in two time periods: 2003 till 2007 and 2008 till 2010. We run four regressions: two for the developing countries and two for the developed countries. The results are presented in Table 5.4.
23 | P a g e Chinese OFDI. The variable Efficiency is also significant, but the coefficient is a small number and therefore relatively not very important. During the economic crisis, none of these variables were significant. This is remarkable, apparently other factors determined Chinese OFDI in this crisis period 2008 – 2010. We need to mention that the time period is short and thus not many observations are included. This could give biased results.
Table 5.4: Determinants of Chinese foreign direct investment, effects of financial crisis, fixed effects estimation
Variables Developing ‘03/’07 Developing ‘08/’10 Developed ‘03/’07 Developed ‘08/’10
Export .18543 * .05868 .01014 .09576 * (.07346) (.03938) (.00507) (.02786) Resources .18399 *** .01929 -.02321 -.53670 * (.01177) (.03127) (.04408) (.17348) Technology .09677 .13278 -.01812 ** .02611 ** (.05784) (.06964) (.00539) (.00468) Efficiency .00029 ** .00007 .00024 *** .00008 (.00009) (.00006) (.00003) (.00005) Political .13650 *** .01580 -.02965 -.08325 ** (.02378) (.05454) (.03038) (.01818) Corruption -.00552 .00260 .01230 .00387 (.00760) (.02221) (.00664) (.02504) Within R² .479 .118 .475 .207 Observations 94 57 68 43
The ‘developing’ and ‘developed’ columns, respectively, provide results based on data from developing and developed countries. Driscoll & Kraay (1998) standard errors are in parentheses. ***, **, and * denote significance at the 1, 5, and 10% levels, respectively.
For the whole time period, the developed countries sample gave results that are consistent with the literature. When we split the time period in a pre-crisis period and during the crisis time period, the results are less clear. Before the crisis, only the variable Technology is (again) negatively significant (Efficiency is also significant but the influence is low). This implies that before the crisis, developed countries probably were protecting their domestic market. As with the crisis period in developing countries, apparently other motives not included in this study determine Chinese OFDI in developed countries before the crisis. During the crisis, we see that Chinese firms invest with a market-seeking motive. Surprisingly, Technology now becomes positive significant. So during the crisis period, Chinese firms invest with a strategic asset-seeking motive in developed countries.
6. Conclusion
resource-24 | P a g e seeking, strategic asset-seeking, and efficiency-seeking motivations. In this paper, we used a fixed effects estimation to analyse the location determinants of the host country that attract Chinese OFDI. Interesting is the fact that more and more Chinese OFDI is invested in developed countries. Thus, we analysed the difference of investment motivations between developing and developed countries.
Our results show that market-seeking and efficiency-seeking investments are motives to invest in developed countries. Developed countries with high-technology industries protect themselves from Chinese OFDI. Market-seeking and resource-seeking investments are motives to invest in developing countries. This has implications for countries that want to attract Chinese OFDI. For example, an efficient work force attracts Chinese OFDI and high imports from China also attract Chinese OFDI. For countries that want to protect their technology advantage, strict regulations could help keeping Chinese firms out of the country. This could also be an example to developing countries that want to keep control over their natural resources.
The financial crisis influenced the investment motives in developing countries. Probably, motives that are not included in this study now influenced the decision to invest or not in a developing country. An example could be the exchange rate between the host countries currency and the Chinese Yen. This exchange rate could fluctuate more in economic downturn and could harm the investment in a specific country.
25 | P a g e References:
Amighini, A., Rabellotti, R., & Sanfilippo, M. (2011). China's outward FDI: An industry-level analysis of host country determinants (No. 3688). CESifo working paper: Empirical and Theoretical
Methods.
Anwar, A. I., Hasse, R., & Rabbi, F. (2008). Location Determinants of Indian Outward Foreign Direct Investment: How Multinationals Choose their Investments Destinations. In CBS Conference on Emerging Multinationals.
Asiedu, E. (2002). On the determinants of foreign direct investment to developing countries: is Africa different? World development, 30(1), 107-119.
Böwer, U., Fernandez D. & Thimann, C. (2009) Outward FDI from Emerging Markets: Evidence from Firm-Level M&A Data. Working paper, European Central Bank
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. Cai, K. G. (1999). Outward foreign direct investment: A novel dimension of China's integration into the
regional and global economy. The China Quarterly, 160, 856-880.
Cheung, Y. W., & Qian, X. (2009). Empirics of China's outward direct investment. Pacific Economic Review, 14(3), 312-341.
Deng, P. (2004). Outward investment by Chinese MNCs: Motivations and implications. Business Horizons, 47(3), 8-16.
Driscoll, J. C., & Kraay, A. C. (1998). Consistent covariance matrix estimation with spatially dependent panel data. Review of economics and statistics, 80(4), 549-560.
Dunning, J. H. and Lundan, S. M., 2008. Multinational Enterprises and the Global Economy, MPG Books Ltd, Second Edition.
Hill, R. Carter., Griffiths, W. E. and Lim, G. C. Principles of econometrics / R. Carter Hill, William E. Griffiths, Guay C. Lim, Fourth edition, Wiley Hoboken, NJ 2011
26 | P a g e Hoechle, D. (2007). Robust standard errors for panel regressions with cross-sectional dependence. Stata
Journal, 7(3), 281.
Kolstad, I., & Wiig, A. (2012). What determines Chinese outward FDI?. Journal of World Business, 47(1), 26-34.
Kyrkilis, D., & Pantelidis, P. (2003). Macroeconomic determinants of outward foreign direct investment. International Journal of Social Economics, 30(7), 827-836.
Liu, X., Buck, T., & Shu, C. (2005). Chinese economic development, the next stage: outward FDI? International Business Review, 14(1), 97-115.
Sauvant, K. P., Maschek, W. A., & McAllister, G. A. (Eds.). (2010). Foreign direct investments from emerging markets: the challenges ahead. Palgrave Macmillan.
Schüller, M., & Turner, A. (2005). Global ambitions: Chinese companies spread their wings. China aktuell, 4(2005), 3-14.
Singh, H., & Jun, K. (1995). Some new evidence on determinants of foreign direct investment in developing countries. World Bank Policy Research Working Paper, (1531).
Stein, E., & Daude, C. (2001). Institutions, integration and the location of foreign direct investment. New Horizons for Foreign Direct Investment, 101-28.
UNCTAD (2011). World Investment Report 2011. Non-Equity Modes of International Production and Development. Geneva: UN.
27 | P a g e Appendix A:
Emerging countries according to the International Monetary Fund:
Argentina Indonesia Poland
Brazil Latvia Romania
Bulgaria Lithuania Russia
Chile Malaysia South Africa
China (excluding Taiwan) Mexico Thailand
Colombia Pakistan Turkey
Estonia Peru Ukraine
Hungary Philippines Venezuela
India
Appendix B:
Developed countries:
Australia Finland Japan South Korea
Austria France Luxembourg Spain
Belgium Germany Malta Sweden
Brunei Greece Netherlands Switzerland
Canada Iceland New Zealand Taiwan
Cyprus Ireland Norway United Kingdom
Czech Rep. Israel Singapore United States
Denmark Italy Slovakia
Appendix C:
Investment (in US dollars) in Asian countries with a resource seeking motive:
28 | P a g e Korea Rep (0.6 billion) Turkmenistan (0.7 billion)
Laos (0.8 billion) UAE (0.8 billion)
Malaysia (0.7 billion) Vietnam (1.0 billion) Some examples:
Investment in the textile industry and energy sector in Cambodia:
http://www.eastasiaforum.org/2013/07/16/chinese-investment-and-aid-in-cambodia-a-controversial-affair/ Seen: 21/04/2014 at 14:00
Investment in the energy sector in Laos:
http://www.eastasiaforum.org/2013/05/25/big-money-big-dams-large-scale-chinese-investment-in-laos/ Seen 21/04/2014 at 14:05
Appendix D:
Table A: Summary statistics whole sample
Variable Obs Mean Std. Dev. Min Max
LnOfdi 262 2.85 2.29 -4.61 7.80 Export 280 13.77 30.07 .01 235.77 GDP 280 1165.70 2389.93 2.73 14958.3 Resources 280 26.61 28.21 .64 98.63 Technology 280 29.44 45.51 .001 220.88 Efficiency 280 16386.45 18359.04 -4073.56 65936.66 Politics 280 73.34 8.58 52.65 91.10 Corruption 280 52.62 20.34 16.67 91.67 Distance 280 8686.09 3774.78 954 16937
Table B: Summary statistics developing countries
Variable Obs Mean Std. Dev. Min Max
29 | P a g e
Table C: Summary statistics developed countries
Variable Obs Mean Std. Dev. Min Max
LnOfdi 111 3.22 2.33 -4.61 7.80 Export 120 26.48 42.33 .48 235.77 GDP 120 2278.96 3307.75 87.42 14958.30 Resources 120 11.42 12.72 .64 64.63 Technology 120 59.19 54.96 .47 220.88 Efficiency 120 35753.05 10491.00 12177.50 65936.66 Politics 120 76.51 6.19 59.85 91.10 Corruption 120 70.00 15.26 33.33 91.67 Distance 120 7766.80 2747.38 954 11156
Table D: Correlation matrix whole sample
Variable LnOfdi Export Resourc. Techno. Efficien. Politics Corr. Dist.
LnOfdi 1 Export .370 1 Resources .260 -.212 1 Technology .273 .737 -.413 1 Efficiency .127 .390 -.392 .558 1 Politics -.048 .064 -.078 .118 .363 1 Corruption .073 .213 -.424 .383 .726 .258 1 Distance -.230 -.135 .097 -.253 -.148 -.081 .065 1
Table E: Correlation matrix developing countries
Variable LnOfdi Export Resourc. Techno. Efficien. Politics Corr. Dist.
LnOfdi 1 Export .463 1 Resources .385 -.055 1 Technology .097 .432 -.334 1 Efficiency .061 .283 .194 .173 1 Politics -.019 -.270 .119 -.096 .213 1 Corruption -.164 -.104 -.311 -.126 -.208 .006 1 Distance -.300 -.387 -.053 -.291 -.078 .020 .285 1
Table F: Correlation matrix developed countries
Variable LnOfdi Export Resourc. Techno. Efficien. Politics Corr. Dist.
30 | P a g e
Table G: Hausman test whole sample: FE vs RE
---- Coefficients ----
Fixed Random Difference S.E.
Export .01959 .02772 -.00813 .00413 Resources .16204 .04891 .11313 .02414 Technology -.00488 .00357 -.00845 .01048 Efficiency .00007 .00001 .00006 .00003 Political -.02730 -.03908 .01178 .01752 Corruption .02953 .02914 .00039 .01141 Chi2 (6) = 41.29 Prob > chi2 = 0.000
Table H: Hausman test developing countries: FE vs RE
---- Coefficients ----
Fixed Random Difference S.E.
Export .13862 .21242 -.07380 .02207 Resources .17516 .03963 .13553 .02778 Technology .09099 .01936 .07163 .06095 Efficiency .00003 -.00013 .00017 .00010 Political .06326 .03443 .02884 .02699 Corruption -.00232 .00675 -.00907 .01382 Chi2 (6) = 27.16 Prob > chi2 = 0.000
Table I: Hausman test developed countries: FE vs RE
---- Coefficients ----
Fixed Random Difference S.E.
31 | P a g e
Table J: Breusch-Pagan Lagrange multiplier test, whole sample: RE vs Pooled OLS
LnOfdi [country, t] = xb + u [country] + e [country, t]
Estimated results: Var sd = sqrt (Var lnofdi 5.261 2.294 e 2.421 1.556 u 1.112 1.054
Test: Var (u) = 0
Chibar2 = 48.25 Prob > chibar2 = 0.000
Table K: Wooldridge test, (1) whole sample, (2) developing countries
(1) Wooldridge test for autocorrelation in panel data (2) Wooldridge test for autocorrelation in panel data F ( 1, 34) = 4.953 F ( 1, 19) = 9.668
Prob > F = 0.033 Prob > F = 0.006
Table L: Modified Wald test, (1) whole sample, (2) developing countries
(1) Modified Wald test for groupwise heteroskedasticity (2) Modified Wald test for groupwise heteroskedasticity in fixed effects regression model in fixed effects regression model
Chi2 (35) = 9716.82 Chi2 (20) = 1708.20
Prob>chi2 = 0.000 Prob>chi2 = 0.000
Table M: Pesaran’s test, (1) whole sample, (2) developed countries
(1) Pesaran’s test of cross sectional independence = 9.029, Pr = 0.000 Average absolute value of the off-diagonal elements = 0.399
