The relationship between
exchange rate and FDI in China
International Economics and Business Faculty of Economics and Management RijksUniversiteit Groningen
Master Thesis By Q. Yan
Abstract:
This paper stands on the FDI host country point, tested how the relative exchange rate change, the relative company wealth in investor country, the relative Ownership advantage and location advantage (cost advantage), and the trade openness affect the FDI inflow. Take China as the host country, this paper observed the foreign investments since 1995 till 2007. The result shows there is indeed strong support for some of the assumptions while does not support others.
Key words:
FDI, exchange rate, relative wealth, ownership advantage, openness, China
*The author gratefully acknowledges Dr. Brouwer, Prof. Dietzenbacher, and PhD
Table of Contents Abstract
1. Problem statement---4
2. Why chose China as the FDI host country---5
3. Literature review and hypothesis---9
3.1. Host country exchange rate appreciation negatively affects FDI inflow ----10
3.2. Investors’ relative wealth positively affects foreign investments---11
3.3. Ownership and Location advantage is the investment motivation ---13
3.4. Openness and FDI bilateral stimulates---16
4. Methodology ---17 4.1. Measurement of variables---17 4.2. Data---20 4.3. Model specification---24 4.4 Model testing ---26 5. Results ---31
6. Discussion and Conclusion---33
6.1 Yearly exchange rate change, ownership and location advantage ---33
6.2 Relative wealth effects for multinationals ---34
6.3 Trade openness, geography location, culture and China economy growth -35 7. Conclusion---36
7.1 Conclusions ---36
7.2 Critics---37
1. Problem statement
Globalization and open markets increased the trades and financial activity across the world. Multinationals (MNEs) invest in other countries for production and financial activities, the result is that both MNEs and countries benefit from foreign direct investment (FDI) and trade (Navaretti & Venables 2006).
There are two main streams of FDI, namely horizontal FDI and vertical FDI. Most of the investments that focus on local market are horizontal FDI and are mainly between developed countries. On the other hand, the export oriented investments are those in the lower level of the production chain, low technology contents which are considered as vertical FDI, which mainly occurs from developed countries to developing countries (Navaretti & Venables 2006). From company level, the foreign investments decisions are not only determined by production resources but MNEs also consider their global strategy and transaction cost reduction (Blonigen, 2005). For example, in order to cover the worldwide market, MNEs build subsidiaries in both developing and developed countries (Blonigen, 2005). To build up the competitive advantage, MNEs establish research and development (R&D) facilities in places where enjoys high productivity, but build production facilities in places where the production cost is low (Navaretti & Venables 2006).
perspectives. Other factors such as tax policy, openness and market size all shows relationship with the FDI flow (Xing, 2006). Thus the determining factors of FDI are not only the exchange rate change, but also the market and resources of the host country, the ownership and the multinational’s global wide strategy (Blonigen, 2005). All the above mentioned factors have been proved to affect the FDI level. To specify the scope of research, this paper will mainly discuss the effect of nominal exchange rate changes on FDI, based on example of a developing country: China.
Among the most common cited factors in academic papers, this paper will test how some selected variables affect FDI. Those variables are: the nominal exchange rate change, the relative labour cost between host country and investor country, the relative price between host country and investor country, stock index ratio and openness. Using the method of multivariate regression, the significance and scope of the relation between FDI and independent variables will be shown. In the next section the literature and choice for these determinants will be elaborated.
2. Why chose China as the FDI host country:
Figure 1. Yearly changes of FDI to China from 1996-2006
Source: China Ministry of Commerce data base.
Figure 2. Chinese currency exchange rate against various currencies from 1995-2006. Measured by percentage change. Source: State Administrative of Foreign Exchange in China
(China custom statistics, 2006). Under the new regime, exchange rate is allowed to float in a managed threshold, so the high pressure from trade surplus pushed up appreciation of RMB against major trade partner’s currency, also leads to the decline of their investment in China. A reason to explain this change is that in 2001 China joined WTO, which stimulates export, leading to an increase of FDI. As the exchange rate to USD kept pegged during 2001 and 2005, the only effect is to increase FDI. Then after the exchange rate reform to a managed floating regime, the pressure of trade surplus leads to RMB appreciation, thus the FDI decreased. We can assume that the FDI in china before 2005 is mainly export oriented, vertical FDI. Exchange rate has an effect to the level of FDI in China, which cannot be ignored especially after the reform of exchange rate regime in 2005.
the ownership reform act in China makes it possible for foreigner companies to purchase state-owned enterprise (Wu, 2003). This paper will observe the scope and extent of horizontal FDI in China for the observed period so that to provide another study of the trend in market orientation FDI.
Lastly, many previous researches focused on the FDI between developed countries, but little research has focused on developing countries as host country for FDI (Blonigen, 2005). This paper will provide a research of FDI in China, which is considered as the largest developing country.
3. Literature review and hypotheses:
Several studies have established the relationship between the nominal exchange rate and FDI decisions. However, the exchange rate is not the only factor that affects the FDI. This section will introduce researches and theories on FDI determinants. Based on previous research, the key parts of this section will analyses the factors which are most likely to affect FDI. Hypothesis will be raised after discussion.
Among many articles study on the FDI determinants; one stream suggests company internal factors or firm-specific factors that motive firms to invest abroad. One representative theory is Dunning’s Ownership-Location-Labour (OLI) theory. OLI theory stresses multinationals invest abroad pursuing comparative advantage. In addition to the OLI theory, Chen (2006) and Blonigen (1997) argue about the company motivation theory. Chen argues that horizontal or vertical production represents the low cost orientation or local market orientation of the investing firm, and Blonigen argues that the acquisition of transferable assets is the motivation of foreign companies. These OLI theory and company motivation theory all suggests that MNEs investing abroad for the host country resources and the market size, in order to achieve their competitive advantage.
The above mentioned theories view the determinants of FDI from different aspects. However, a clear message is that there are many factors affecting FDI. These factors can be inter-related or co-related, determinants of FDI is not based on single factors but the interaction of various factors. This paper will focus on the country level of FDI however will look into the company internal factors that affect the investment abroad and also the external factors in destination country that affect on FDI. In the following chapters will explain the variables in more detail.
3.1. Host country exchange rate appreciation negatively affects FDI inflow:
Several studies have established the relationship between the nominal exchange rate and FDI decisions. First we want to find what are the aims and sources of FDI, and then we will discuss the exchange rate and FDI relationship.
Another view is “market oriented” FDI. Kohlhagen (1977) argues that domestic firms would defer foreign direct investments until after the foreign currency devaluation, and then it would be more profitable relative to exporting. This means that if a MNE aims at the market of that country, it will invest into that country after the currency depreciation in that country. In general, no matter it is market oriented or exports oriented FDI, foreign investments are likely to be increase if the host country currency depreciates.
Concerning the different types of foreign investment, Goldberg and Kolstad (1995) suggest that most FDI are production and equity investments. They argue that the key to FDI is ownership. FDI largely consists of equity and debt transfer, merger and acquisitions by firms in affiliated corporations located in countries other than the home country of the investor. Because a large components of FDI are foreign merger and acquisitions (Dewenter, 1995), we then assume that FDI is as sensitive as exchange rate changes. Harris and Ravenscraft (1991) and Swenson (1993) have provided that depreciating dollar will bring higher foreign take over transactions and also higher flows of FDI. To end up the arguments, we propose as follows
Hypothesis 1: Chinese currency appreciation reduced FDI in China, and vice versa, between 1995 and 2007.
3.2. Investors’ relative wealth positively affects foreign investments:
Froot and Stein (1991) introduced the imperfect capital market theory, which states that the capital market is information asymmetric and not a free market. Froot and Stein argue further that the internal cost of capital is lower than external cost of capital. The consequence of this imperfect market theory is that firms with higher wealth can have higher ability to financing and acquire assets. The ‘wealth effect’ measures the relative wealth of foreign investors to domestic investors.
encouraging foreign funds to acquire more in order to generate more return (Xing, 2006). This means that the depreciation of host country currency will increase the investing company’s wealth in terms of host country currency, and decrease the real wealth of host country firms, thus foreign investors will have more funds to invest relative to the counterpart firms in hosting country, encouraging foreign investors to acquire more assets (Froot and Stein, 1991). This is the reason why a depreciation of host country currency will increase the inward FDI in the host country. The higher the ratio of the foreign currency versus host country exchange rate, the higher will be inward FDI in the host country (Liu et al. 1997).
If the wealth effect is largely driven by the exchange rate changes, a depreciation of Chinese currency will increase relative wealth of foreign company and thus increase FDI. In addition, foreign investors will be attracted by the relatively cheaper labour costs when Chinese currency depreciates; this has been discussed in Hypothesis 1: Chinese currency appreciation reduces FDI inflow.
Increase in multinational’s market value also increases the company wealth, and further increase their domestic expansion as well as foreign expansion. Market value is the value of the company when it is sold or being traded. The market value can be measured by the stock of the company that listed on stock market. An increase in investors’ company wealth stimulates them to expand and acquire more assets (Backer et al, 2008).
consequence will be that both the country’s investment outflow and foreign stock market value increases. However it is hard to track the result of host country stock market value increase as a result of FDI inflow. In this paper we suppose that when the stock market index in investors’ country increases, the FDI from that country also increases (Backer et al, 2008).
Increased company wealth will stimulate their domestic or foreign expansion. The investor country’s stock market should demonstrate a positive relationship with the FDI outflow (Baker et al, 2008). We expect foreign companies will increase investments in China by acquiring Chinese companies or increased investments. Therefore, we expect the positive relationship between the investors’ home country stock market growth (firm value) and their investments abroad. Then in hypothesis 2 we assume the following:
Hypothesis 2: company wealth growth measured by country’s stock market index growth has positive relationship with FDI inflow in China in 1995-2007.
3.3. Ownership and Location advantage is the investment motivation:
companies that have such local experience or with a certain size. In addition, foreign investors often are considered as some that have stronger financial ability and advanced management skills enabling them perform better than Chinese local counterparts. Dunning argues that the greater the ownership specific advantages of the investing firms comparing with other firms of the destination country of investment, the more they are likely to invest in that country (Dunning, 2000). The ownership advantage comes from the benefit to the investors after acquisition of foreign assets. For example, foreign investors can build up global economic of scope by acquiring Chinese assets, while Chinese counterparts might not achieve because of lacking knowledge or technology. As argued by Eiteman et al. (2004) MNCs invested in China have competitive advantages against Chinese local firms in terms of economies of scale and scope, managerial and marketing expertise, technology, R&D skills, financial strength and differentiated products. With these advantages, they can easily defeat Chinese local companies and gain large profit. Those companies depending on their specific advantages are moving aggressively into China, not only to take a temporary benefit but also for the long haul of the market potential (Liu et al, 1997).
Most of the investors in China come from developed country, those countries have relatively higher productivity, GDP and GDP per capita ratio than China, and their companies are considered as strong in technology, global network and financing (Phelps, 2008). As a result, acquisition of Chinese assets can strengthen their global profitability and companywide knowledge accumulation, enabling them to better compete with rivals in China and abroad. Because of knowledge accumulation and investment revenue maximisation, recently more and more multinationals prefer wholly owned investments in China. From this aspect, we can say that ownership specific competitive advantages, measured by high country productivity, are one of the important determinants of foreign firms’ investment in China; so that we suppose the greater those advantages, the higher the level of FDI.
Another sub-paradigm is the location attractiveness of hosting countries. The location paradigms are such as the country endowments or natural resource (Blonigen, 2005). Examples of location-specific factors are markets, resources, production costs, and cultural affinities (Dunning, 1977). This sub-paradigm avers that the more the immobile, natural or created endowments, which firms need to use jointly with their own competitive advantages, favor a presence in a foreign, rather than a domestic location, the more firms will choose to pursue or exploit their O specific advantages by engaging in FDI (Dunning, 2000). In other words, multinationals will invest in countries that have location advantages in order to strengthen their O specific advantage. In general, investing in location where resources and endowments exist can strengthen their Ownership advantages.
Since the 1980s the globalization process increased, it’s more and more important for those firms to establish affiliates abroad to achieve economy of scale and scope, production efficiency and share comparative advantages. In China, the labor resources and infrastructure become the most important comparative advantages both in the long-run and short-run. Nowadays accompany with the Chinese economic growth, labour productivity also increased significantly (Xing 2005, Zhang et al 2007), so not only the cheap labour cost attracts MNEs, but also the increased labour productivity increased the location attractiveness for the multinationals.
Hypothesis 4: the higher the country’s relative Location specific advantage against China, measured by relative labor cost of investing country to China, the higher the level of FDI from that country will be in the period of 1995-2007.
3.4. Openness and FDI bilateral stimulates:
Openness includes trade openness and financial openness as stated in WTO complement agreements. China has been liberalized the financial market since 2006 as the implementation of the WTO agreement (Xing 2006). This section will discuss the trade perspective. Overall, there are two different perspectives with respect to the impact of trade openness on FDI or the relation between trade and FDI. The first perspective is developed by Mundell (1957) who suggested that there is a substitution relation between trade and FDI. In his model, Mundell states that firms invest abroad or engage in FDI because they want to get rid of the high trade barrier. This implies that the higher the barrier or the lower the trade openness, the lower the amount of trade and the higher the amount of FDI which substitute the reduced trade. The other perspective is the mutual complement theory of trade and FDI. This perspective is opposed to the substitution theory of Mundell, suggesting that reasonable foreign investment will improve trade and under certain condition, trade and FDI will mutually encourage and improve each other (Markuson and Svensson, 1985; Kiyoshi, 1987).
Both kinds of theories are valuable and are suitable for specific situations. Mundell’s theory is more appropriate for market-oriented FDI, while the mutual complement theory fits export-oriented FDI better. With regard to the condition of FDI in China, the second theory works better because inward FDI in China is deemed to be more export oriented (Zhang et. al. 2007).
suggests trade and investment mutual complement theory. We assume the higher the foreign trade with china is, the higher the FDI in China, resulting in:
Hypothesis 5: the higher the level of trade openness of China to the investing country, the higher the level of FDI from that country is, in the period of 1995-2007.
4. Research methodology 4.1. Measurement of variables:
The aim of this paper is to test the effect of the nominal exchange rate change and the wealth effect on determinants of FDI in China. There are many variables that can affect the FDI inflow over time, as discussed in previous chapter, there are company internal variables and external variables, however, due to data availability and time limitations, some variables cannot be included in this research. For example the changes in tax policy, the inflation rate, food and agriculture price, real estate price and the government reform policy will exempted from this analysis.
To operationalize the measurement, we will use the following formulas, ‘it’ represents the measurement of investing country ‘i’ in year t; similarly ‘ct’ represent the measurement of China in year‘t’.
With regard to the exchange rate change, we use nominal exchange rate change and which is the focus of this study. A widely used formula is employed to calculate the yearly percentage change (depreciation or appreciation of foreign currency against RMB under a direct quotation scheme, which is RMB value per foreign currency). This kind of percentage change of exchange rate is calculated as a percentage increase or decrease of exchange rate from the beginning of the year to the end of the year. Beginning of the year is January and December is the end of the year. “t” represents the year of measurement. The formula is shown in formula 1. The regression result is expected to be significantly negatively related to FDI, the reason is that Chinese currency appreciation indirectly means foreign currency depreciation, that will increase the investment cost so that reduce their investment in China.
EXit= (1)
To measure relative company wealth effect, we use the cumulative wealth of the country’s firm. Stock market index measure the relative company value of the country. The foreign stock market index against Chinese stock market index shows the ratio of relatively wealth. For each country the stock index is in term of their own currency, the country’s company wealth change is measured in percentage of stock index changes, as denoted in formula 2. The test result is expected to be strong positive related to FDI, as the higher foreign company wealth, the more ability they expand internationally.
STOCKit = (2)
managerial skills between foreign investor and Chinese local firms, so the country productivity and country income is a good measurement. GDP per capita is deemed as a predictor of the development and technology level as well as productivity level of a country. It is reasonable to assume that firms from countries with higher GDP per capita have more owner specific advantages. In addition, relative GDP of investing countries to China can also be considered as a measurement of this variable because it represent the number of large firms with owner specific advantages in a country (Ajami and Barniv, 1984). Furthermore, GDP ratio is a direct indicator of the economy size between two countries, thus indirectly represents the relative market size between two countries. GDP means the nominal GDP because the exchange rate, stock value and labour cost are measured in nominal term. The relative GDP ratio and relative GDP per capita ratio are calculated through the following Formulas 3 and 4 respectively. Nominal GDP is measured in terms of U.S. dollar; both GDPR and GDPPC are taking the ratio between foreign countries against Chinese value so the ratio is in terms of percentage. The regression result is expected to be positive relationship for both these two variables. The reason is that the higher the multinational advantage, measured by the sourcing country productivity and technology level, the more they have competition advantage than Chinese companies, so that investing in China can make them win the competition and get higher profit.
GDPRit = (Relative GDP ratio) (3)
GDPPCit = (Relative GDP per capita ratio) (4)
The regression result is expected to be positive, because the higher the labour cost abroad than in China, the more manufactures will move to China seeking cheap labour.
LCit= (5)
(Relative labour compensation)
Trade openness can be measured by the money value of Chinese exports plus imports from the foreign country divided by nominal GDP of that country, which is a widely used measurement of trade openness. Here by adding the GDP into the denominator some bias can be reduced (Yanikkaya, 2003). For example the same amount of total trade amount between a large foreign country and a small foreign country may mean different things. For smaller country it may stand for the high trade dependence on China and high trade openness, while to a large country, the same amount of trade may only take a small percentage of its economy and therefore may mean low trade dependence on China and low trade openness. The import & export money value and the GDP of that foreign country should in same currency term in order to unify the measurement. The relationship is expected to be positive related to FDI because the more China integrates to the world economy, the more trade and also more FDI inflow. Calculation of trade openness can be seen in Formula 6,
TOit = (6)
4.2. Data:
this panel data captures as many of variables that affect each country, and for the same country, it can also capture the changes across time (Harrison, 1994).
Table 1 shows the top 16 regions investing in China since 2002, which are Hong Kong, Japan, Korea, Taiwan, Singapore, US, UK, Germany, France, Canada, Netherlands, Marco and the free harbors ( Virgin islands, Cayman island, Samoa, Moriches). In 2006 the FDI from those regions takes up to 85.6% of total FDI in China. To simplify the analysis, top 10 countries from the list are selected, that is Hong Kong, Japan, Korea, Taiwan, Singapore, US, UK, Germany, France, Canada. These ten countries still take up 60% of total FDI in China in 2006. In this way they may still have good representation of the population of countries investing in China.
2007 2006 2005 2004 2003 2002 1 HK 2770342 2130718 1794879 1899830 1770010 1786093 2 Virgin Isl. 1655244 1167728 902167 673030 - - 3 Korea 367831 399319 516834 624786 448854 272073 4 Japan 358922 475941 652977 545157 505419 419009 5 Singapore 318457 246300 220432 200814 205840 233720 6 US 261623 299995 306123 394095 419851 542392 7 Cayman Isl. 257078 213175 194754 204258 - - 8 Samoa 216988 161977 135187 112885 - - 9 Taiwan 177437 222990 215171 311749 337724 397064 10 Moriches 133250 110551 - - - - 11 UK 83094 75452 96475 79282 74247 89576 12 Germany 73397 200297 153004 105848 85697 92796 13 Marco 63700 67774 60046 54639 - - 14 Netherlands 61666 86460 104358 81056 72549 57175 15 France 45601 39534 61506 65674 60431 57560 16 Canada 39658 44175 45413 61387 56351 58798
Table 1 Non-financial Realized FDI in China (10 thousand $)
Table 2 China foreign reserve during 1978-2006
Source: National statistics. http://www.stats.gov.cn/tjsj/ndsj/2007/indexce.htm
Table 3 Total FDI in China during 1985-2007
Source: China Ministry of commerce. http://www.mofecom.gov.cn
The data of FDI are collected from the China Ministry of Commerce, in term of 10 thousands U.S. dollar. The data are non-financial FDI since the financial market is much more sensitive than production sector, testing the non-financial foreign investment can better reflect the labour productivity and wealth effect that affect the FDI flow. From table 3 we can see that the Chinese FDI inflow follows a steady but rapid growth path, though during the year 1998 and 2000 there is a decline due to the Asia crisis.
The exchange rate is the historical nominal yearly average rate that published by various sources, using direct quotation, taking RMB as term currency. For the dollar, Japanese Yen and Hong Kong dollar rate is available since 1994 from China central bank and the State Administrative of Foreign Exchange. The remaining exchange rate data is collected from American Federal Reserve Bank; currency for like Germany and France has been changed to EURO after 1999, so the exchange rate for these two countries since 1999 is the EURO rate published by European central bank. The exchange rate prior 1999 is the Germany Mark and French Franc taken from central bank of each country respectively.
Korea national statistic office. The stock indices for Hong Kong (HangSeng index), Japan (NIKKEI225 index), Singapore (STRAITS index), U.K (FTSE100 index), France (CAC index) and Germany (DAX index) are available on Yahoo.
Nominal GDP and GDP per capita data are collected from International Monetary Fund (IMF) statistics yearbook and Central Intelligence Agency (CIA) world fact book. The value is in terms of U.S. dollar. Chinese data for the years before 1997 are collected from the China statistics yearbook which published annually by the National statistics bureau, the value were in Chinese currency, we transfer it into U.S. dollar value according to the nominal exchange rate of that year.
The hourly compensation of workers is adopted from the index developed by the US bureau of labour statistics. This data only available till 2004 so in our sample the time period is from 1995 till 2004.
4.3. Model specification:
Statistic data as well as various academic researches has proven that there is significant correlation between FDI and the currency value in U.S. (Klein & Rosengren, 1994; Froot & Stein, 1991). To present the relationship between FDI and the independent variables, Klein and Rosengren (1994) analyzed the U.S. inward FDI from the year 1979 to 1991; they tested the relative wealth, labour cost and the real exchange rate. The result showed strong support for wealth effect above all other effects. Xing (2006) tested the Japanese manufacturing FDI investment in China during the year 1981 and 2002. Xing conducted a multi-regression with real exchange rate, wage and openness as independent variables. The result shows the exchange rate is one of the most significant variables affecting Japanese FDI in China. Both Xing (2006) and Klein and Rosengren (1994) used the panel data and to observe the relationship between independent variables and the dependent variable. Following the previous research, this paper uses the following regression model:
FDI=f (EX, SYOCK, RGDP, RGDPPC, RLC, TO) (7)
change of stock index of each country which reflecting the wealth effect of company market value; RGDP stands for the relative annual GDP of investing country to China; RGDPPC measures the relative annual nominal GDP per capital of investing country to China; RLC is the relative hourly compensation of investing country to China; TO stands for trade openness of China to investing country. Formula (7) can be further expressed as following empirical model:
log(FDI)it=β0+β1xEXit+β2xSTOCKit+β3xRGDPit
+β4×GDPPCit+β5×RLCit+ β6×TOit+ ε (8)
Here i stands for foreign country (i=1, 2, ..., 10), t stands for time series (t=1995, 1996, ... 2007), εit stands for the residual error for the model. FDI is taken in logarithmic form in order to reduce the heteroskedasticity.
β0 is the parameter in the model. There will be two control variables in this model in
order to control for country specific effects. First is the Chinese economy nominal GDP growth rate (CNGDP), the second is whether the foreign country has location specifications (ASIA), which treated as a dummy variable. Firstly, it is worth to see if the China economy growth itself to the attractiveness of Foreign investment. The relationship is expected to be positive because the larger Chinese GDP growth means there is a bigger market so that attract foreign investors to involved in. I include the china nominal GDP growth rate as control variable. The variable is measured as the percentage change of GDP from one year to previous year. That is, calculated as the GDP for the second year minus the first year then divided by the GDP of the first year. The data is as mentioned in previous chapter come from International Monetary Fund (IMF) and China statistics Bureau.
The Eviews software is used to examine the model to see whether those factors have significant contribution to the FDI in to China and what’s the overall power of the model in explaining the variance in FDI.
The model uses panel data because the dataset contains observations on multiple variables observed thought multiple time periods. There are 10 countries in the 13 years period 1995-2007, each country has 6 observations. Given the time series and cross-section characteristics of the data, we have to choose between random or fixed effect models.
The random effects model assumes that the individual country in the sample are randomly chosen and to represent a larger population of countries. In contrast, the fixed effects model assumes that all behavioral differences between individuals and over time are captured by the intercept (hill et al. 2001). In order to choose between fixed or random effects model, Hausman test will be used. Hausman test compares null hypothesis of coefficients estimated by random effects estimator are the same as the ones estimated by the fixed effects estimator. If in the test the residual shows a significant p-value, a fixed effects model should be used, and vice versa. Generally speaking, fixed effects model is commonly used in panel data analysis (Xing, 2006). In this paper I suppose that both estimation equations display a significant p-value of the residual. In our model, the Chinese GDP growth rate is the same for each country where the FDI comes from. There fore this variable is considered as fixed variable, which explains the use of fixed effects model. Finally, after test adding or remove control variable, the model result (table 9 and 10) display a significant p-value. Thus this paper employs the fixed effects model.
4.4 Model testing
In this section the dataset will be tested to see whether it fits with multiple regression model (Hill et al., 2001). The testing methods are discussed below and the results will be discussed later.
correlation between two variables is higher than 0.8, there might be multicollinearity. In that case, the variables doesn’t sufficiently estimate the separate effect on dependent variable because the p-value is higher even these two variables are significant determinants of dependant variable (Hill et al. 2001).
Stationary time series measures whether the model is correct. A Unit-Root test can be applied taking the null hypothesis of a Unit Root. If there is non-stationary series, we have to take the first difference to arrive at stationary time series. Homoskedasticity measures all observations have equal variances; it can be tested by using the White heteroskedasticity test with the null hypothesis of equal variances. Autocorrelation implies that the current error term is also affected by shocks from previous periods. Durbin-Watson test is used to observe the autocorrelation. Finally the residuals need to be normally distributed since the hypothesis tests rely on the assumption of normal distribution of the errors and the dependent variable. Therefore the Jarque-Bera test carrying the null hypothesis of errors being normally distributed needs to be run (Hill et al. 2001).
The model testing results are shown below. Firstly, the descriptive statistics (table 4) shows a low standard deviation, resulting from a low spread of the values in the dataset. The average value of RGDPPC is relatively high compared to other variables; however it is acceptable as the GDP per capita might varies across countries according to their population.
LOGFDI EX STOCK GDPR RGDPPC RLC TO CNGDP ASIA Mean 2.940071 -0.003801 0.086027 1.785195 25.87970 21.34802 0.016843 9.086301 0.503876 Median 3.063391 -0.002559 0.096477 0.808962 24.36226 19.07631 0.012027 0.128727 1.000000 Maximum 5.331752 0.403671 0.784188 10.48043 72.68870 65.54307 0.067859 116.5373 1.000000 Minimum 0.955511 -0.478283 -0.517081 0.048368 6.666019 4.419890 0.001762 0.062496 0.000000 Std. Dev. 1.122233 0.095800 0.236317 2.464509 11.39204 12.49403 0.015415 31.13847 0.501934 Skewness 0.291047 -0.143708 0.026442 2.008498 0.806404 0.910314 1.358095 3.175412 -0.015504 Kurtosis 2.512476 8.745675 3.367626 6.097023 4.178653 3.833894 4.126847 11.08328 1.000240 Jarque-Bera 3.122780 179.2667 0.741456 139.3589 21.61453 16.70860 46.84050 572.3917 21.50000 Probability 0.209844 0.000000 0.690232 0.000000 0.000020 0.000235 0.000000 0.000000 0.000021 Observations 130 130 129 130 130 100 130 130 129 Table 4: Descriptive statistics
to take into account the fact that due to the relative high correlation between variables their individual relationship towards our dependant variable can not be fully explained by the model.
CNGDP ASIA EX GDPR RGDPPC RLC TO STOCK CNGDP 1.000000 -0.024447 0.017389 0.084392 0.288036 0.214224 0.051591 0.005973 ASIA -0.024447 1.000000 -0.087003 -0.342956 -0.265934 -0.599373 0.654565 -0.132811 EX 0.017389 -0.087003 1.000000 -0.033909 -0.140671 -0.122221 0.038811 0.238520 GDPR 0.084392 -0.342956 -0.033909 1.000000 0.606791 0.523926 -0.297061 0.033538 RGDPPC 0.288036 -0.265934 -0.140671 0.606791 1.000000 0.744686 -0.315389 0.078598 RLC 0.214224 -0.599373 -0.122221 0.523926 0.744686 1.000000 -0.585939 0.137515 TO 0.051591 0.654565 0.038811 -0.297061 -0.315389 -0.585939 1.000000 -0.017189 STOCK 0.005973 -0.132811 0.238520 0.033538 0.078598 0.137515 -0.017189 1.000000 Table 5: correlation matrix
The result of Unit- Root test shows a rejection of null hypothesis of unit root in the series. For example, Dicky-Fuller method (ADF) in Unit-root test shows a significant rejection of null hypothesis. We can conclude that this is stationary time series (table 6).
Group unit root test: Summary Sample: 1 130
Method Statistic Prob.** Cross-sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -5.18859 0.0000 6 770 Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -11.9382 0.0000 6 770 ADF - Fisher Chi-square 173.411 0.0000 6 770 PP - Fisher Chi-square 189.830 0.0000 6 772 Table 6: Unit-Root test for stationary.
Heteroskedasticity Test: White
F-statistic 1.113915 Prob. F(27,71) 0.3497
Obs*R-squared 29.45807 Prob. Chi-Square(27) 0.3390 Scaled explained SS 25.14397 Prob. Chi-Square(27) 0.5664 R-squared 0.297556 Mean dependent var 0.140327 Adjusted R-squared 0.030430 S.D. dependent var 0.198300 S.E. of regression 0.195260 Akaike info criterion -0.195756 Sum squared resid 2.706971 Schwarz criterion 0.538217 Log likelihood 37.68991 Hannan-Quinn criter. 0.101211 F-statistic 1.113915 Durbin-Watson stat 1.441660
Prob(F-statistic) 0.349690
White heteroskedasticity test result shows the P-value of F-stat is 0.3497, meaning that cannot reject the null hypothesis. The null hypothesis of the White test is “no heteroskedasticity”, so we conclude that there is no heteroskedasticity in the model (table 7).
The P-value of Jarque statistic is 0.925 (table 8), meaning cannot reject the null hypothesis. The null hypothesis is normally distributed error terms. So we accept normal distribution is the error terms.
Table 8: Jarque-Beta test result for residuals
The autocorrelation Durbin-Watson result (table 9) shows value of 0.6, so that we can not reject the null hypothesis. The null hypothesis assumes that there is no autocorrelation. So we conclude that there is no autocorrelation. To improve the Durbin-Watson result and to reduce autocorrelation, several methods can be used, such as increase the time period of observation so that increase the length of panel; increase the number of repressors; and increase the number of individuals in the panel.
5. Results:
From the result table (table 9) the F-statistic (R-square=0.886) which means nearly 88.6% of the variation in total FDI is explained by the variation in independent variables together. Leave around 12% of the observations is left unexplained and is due to variation in the error term or to variation in other variables that implicitly from part of the error term. With 5% significance level, most of the independent variables present significant result that explains the dependent variable except the STOCK variable. The variable for China exchange rate yearly change (EX) towards foreign currency has negative but significance effect to the FDI inflow (0.0075). That proves Hypothesis 1 that Chinese currency appreciation will reduce FDI inflow.
Dependent Variable: LOGFDI Method: Least Squares
Coefficient Std. Error t-Statistic Prob.
C 1.849867 0.143463 12.89442 0.0000 EX -1.099423 0.401984 -2.734996 0.0075 STOCK 0.158360 0.162468 0.974713 0.3323 GDPR 0.166134 0.018906 8.787438 0.0000 RGDPPC 0.040750 0.005881 6.929324 0.0000 RLC -0.054835 0.005731 -9.568543 0.0000 TO 53.87574 3.571968 15.08293 0.0000
R-squared 0.885748 Mean dependent var 2.982184
Adjusted R-squared 0.878297 S.D. dependent var 1.113891 S.E. of regression 0.388592 Akaike info criterion 1.015511 Sum squared resid 13.89237 Schwarz criterion 1.199005 Log likelihood -43.26781 Hannan-Quinn criter. 1.089753 F-statistic 118.8727 Durbin-Watson stat 0.604869
Prob(F-statistic) 0.000000
Table 9: Eview regression results fixed effect model. Significant at 5% significance level
Stock index unfortunately has insignificant effect to the FDI to China, though the positive coefficient gives us a signal that there might be positive effect. Hypothesis 2 is rejected as there is no significant effect of multinational wealth, measured by foreign stock index growth, to China FDI inflow.
productivity gap between foreign countries to China, the more they will invest to china pursuing Ownership advantage. Hypothesis 3 is accepted.
Relative labour cost ratio (RLC) shows a negative and significant effect to China FDI inflow. That means the higher salary in foreign country against China, the less FDI. Hypothesis 4 is rejected as the result is negative but significant effect to FDI inflow. Trade openness such as import and export trade has positive and significant influence to the FDI to China. Hypothesis 5 is strongly supported, the more international trade, the more FDI flow to China.
Adding the control variables, namely the China GDP growth rate, regional location has little impact to the panel regression result. As shown in table 10. Though the significance of China exchange rate change has reduced but remains significant (0.0408), still proves that the Chinese exchange rate has strong effect to China’s FDI inflow. Relative wealth indicator STOCK remains insignificant to affect the FDI in China. Chinese economy variable, the Chinese GDP growth shows significant negative effect to FDI inflow. This means the higher Chinese economy growth, the less FDI it will receive. Region and culture variable (ASIA) is a dummy variable, countries in Asia are considered geographically close to China and shares similar culture, this location variable shows significant and positive effect to Chinese FDI inflow, indicating the regional effect in trade and FDI.
Dependent Variable: FDIHUNDREDMILDOLLAR Method: Least Squares
Coefficient Std. Error t-Statistic Prob.
C 1.827708 0.134988 13.53980 0.0000 EX -0.732600 0.353024 -2.075210 0.0408 STOCK 0.165119 0.140777 1.172919 0.2439 GDPR 0.171470 0.016685 10.27667 0.0000 RGDPPC 0.037385 0.005427 6.888223 0.0000 RLC -0.042543 0.005487 -7.753484 0.0000 TO 49.77888 3.467476 14.35594 0.0000 CNGDP -2.468192 0.654636 -3.770328 0.0003 ASIA 0.443244 0.102686 4.316490 0.0000
R-squared 0.916742 Mean dependent var 2.982184
Log likelihood -27.60268 Hannan-Quinn criter. 0.834902 F-statistic 123.8726 Durbin-Watson stat 0.585561
Prob(F-statistic) 0.000000
Table 10: Eview regression results fixed effect model including control variables. Significant at 5% significance level
Worth to note that, in order to seize the impact of China’s WTO membership in 2001 and the financial market openness in 2006, and also taking into account these two single years might affect the panel dataset, I have also tested the years 1995-2000, 2001-2006, and 1995-2006 separately, however, the result does not changes significantly.
6. Analysis and discussion
The result shows significant support for the hypothesis except the company wealth effect hypothesis (stock index of investing country). The control variables do have minor effect to the significant of the variables but the effect can be neglected.
6.1 yearly exchange rate change, ownership and location advantage
Hypothesis 1 is strongly supported by the regression result, meaning that Chinese currency appreciation have negative effect to FDI inflow. This result supports the “export oriented” FDI assumption, as with Chinese currency appreciates, foreign investments declines. The reason might be that RMB appreciation reduces the international competitiveness of Chinese products, reducing the investment return so that multinationals switch to other country where higher cost advantages have. Another reason is that depreciate RMB reduces the cost of investing in China, so that it is more profitable to invest in China when the RMB depreciates.
more and more FDI are focusing on Chinese local market. Higher labour cost gap between foreign countries to China does not attracts higher FDI to China. The conflicting result between the “export oriented” assumption supported by exchange rate changes and the “market oriented” assumption supported by relative labour cost coexists. One explanation is that China is under a transit process that is trying to transform “export oriented” FDI to “market oriented” FDI. Reducing the labour income gap between China and foreign countries will attract more FDI. Recent news proves our assumption: more and more MNEs are planning to set up their R&D center in China this means China is increasing productivity and increasingly important in international economy. (State administration for industry & commerce, 2004)
The relative GDP ratio and relative GDP per capita ratio shows significant effect to FDI to China. These two variable measures the relative technology and productivity advantage for foreign countries against China, which is the “Ownership advantage” of foreign countries to China. As China has been the world’s 7th largest country, developed country such as U.S., and EU countries are likely invest in China. Multinationals from developed countries have technology and skill advantage as well as higher financing ability than Chinese counterparts, multinationals can enjoy competitive advantage and economic of scale than Chinese firms in Chinese market, this comparative advantage attracts multinationals investing in China and enjoys high investment return. This finding supports the OLI assumption.
In addition, the ownership is mostly and can be easily achieved by acquisition of existing assets (Eiteman et al. 2004), as multinationals have stronger financing ability; it is easier for multinationals to purchase Chinese assets than local counterparts. The above description tells us that developed country are more interesting investing in China, and because productivity often positively related to labour cost, the more China increases its productivity, the more FDI inflows.
6.2 relative wealth effects for multinationals
cannot conclude that the company wealth increase in home country will lead multinationals increase their investment to China. One explanation is that international investment is more complex than domestic investment. Multinationals have more to consider before invest to China. For example, the risk of investment is higher abroad than domestic, the regulations or culture might be different in China than home so that multinationals have to solve more problems. In addition, both the productivity and labour cost increases in China, when multinationals making “export oriented” investments they can choose other country such as Mexico and India. On the other hand, “market oriented” investments such as Greenfield investment and acquisitions in China might enjoy the high economic growth and lead to higher investment return, it is a investment stock growth process but not the other way around as we assumed.
Test result for relative wealth effect shows that company market value does not affect the FDI decision to China. However, the test result for exchange rate shows there is significant negative effect of exchange rate. Chinese currency appreciation reduces FDI to China, indirectly indicates that foreign currency appreciation will increases their investment to China. So we can see that the foreign exchange market is relatively sensitive than stock market between China and foreign countries, indicating that the foreign stock market is not directly related to foreign manufacture investments.
6.3 trade openness, geography location, culture and China economy growth
Chinese economy growth as a control variable shows significant positive effect to FDI to China. This means that Chinese growth provides the opportunity for higher investment return for multinationals, the higher economy growth in China, the more attractiveness for China to multinationals, thus attracting more FDI to China.
The geography location as control variable indicates the location and culture closeness have positive and significant effect to FDI in China. This result indirectly shows that regions with similar culture are easily to have intensive business involvement such as cross-country investments. We can also assume that regional integration is more likely to occur between countries that are geographically close and shares similar culture.
7. Conclusion 7.1 Conclusion:
This paper discussed the effect of various factors on FDI inflow to host country, tries to apply the theories (exchange rate theory, OLI theory and the trade openness theory) to the determinant of FDI inflow. Multivariable regression model is used to measure the Chinese FDI inflow from ten major investing countries in the period 1995-2007. This paper trying to find to what extent the variables affect FDI inflow. The variables are: relative exchange rate change between foreign country and host country, the ownership and location advantage between foreign investor and host country, relative labour cost (labour productivity), and trade openness of host country to investor country
China and foreign investor countries will attract more FDI. The explanation is FDI is switching from export oriented investment to market oriented FDI, multinationals increase their investment in china, increases R&D and market penetration investment, which increases Chinese income level as well as labour cost. Most of the investors come from developed countries, multinationals investing in China take ownership advantage by having better management skills, stronger financing ability and economic of scale. The larger the ownership advantages, the more investments into China. Company relative wealth increase measured by the market value does not have impact on their foreign investment. Foreign investment might have high return however associated with higher risk than domestic investment, companies might afraid of foreign expansion.
The result is significant support the exchange rate appreciation reduces FDI, the company ownership positively stimulate FDI. In contrast, reducing labour cost does not attract more FDI but increasing the labour productivity and increasing income will attract more FDI.
International trade and FDI bilateral stimulates, this paper strongly support that trade openness stimulates FDI into China. Increasingly intensive foreign trade increases the market share of Chinese product.
This paper tested the FDI to the biggest developing country—China, where there was little relevant research. This paper tested the joint effect of exchange rate, company internal effect: wealth effect as well as company incentives: ownership or location advantage. This paper can be used as reference for authorities make exchange rate and state-owned asset restructure decision.
7.2 Critics:
Due to data collection limit and time limitation, this paper tested the effect of exchange rate effect, company incentive effect (ownership, location advantage decision), company wealth and trade intensity on effect of FDI to China. There might be variables are not included into this research, so further research can focus on how other external or internal factors such as tax, the regulations affect FDI.
This paper treated company internal factors such as export orientation and low cost orientation together with external factors (i.e. exchange rate, industry) however it might be more focus if only test the external factors because exchange rate and other external factors have more common characteristics so that the analysis is more focused.
8. Reference:
Andersson, J. E., 1979. A Theoretical Foundation for the Gravity Equation, American Economic Review, Vol. 69, issue 1, pp. 106–16.
Aizenman, J., 1992. Exchange Rate Flexibility, Volatility and Patterns of Domestic and Foreign Direct Investment, International Monetary Fund Staff Papers vol.39 no. 4. Pp. 890-922.
Ajami, R. A. and Barniv, R., 1984. Utilizing economic indicators in explaining foreign direct investment in the U.S., Management International Review, vol. 4. pp. 16-26. Baker, M., Foley, C. and Wurgler, J., 2008. Multinationals as Arbitrageurs: The Effect
of Stock Market Valuations on Foreign Direct Investment. Review of Financial Studies, Advance Access published online.
Blonigen, B. A., 1997. Firm-Specific Assets and the Link between Exchange Rates and Foreign Direct Investment. American Economic Review, Vol. 87, pp. 447-465. Blonigen, B. A. 2005. A review of the empirical literature on FDI determinants.
Atlantic Economic Journal. Volume 33, Number 4.
CA S&PTS Index report 2005. 2005. Toronto Stock Exchange. Toronto CA 2005. Carr, D. L.; Markusen, J. R.; Maskus, K. E., 2001. Estimating the Knowledge-Capital
Model of the Multinational Enterprise. American Economic Review. vol. 91, issue 3, pp. 693-708.
Caves, Richard E. 1996. Multinational Enterprise and Economic Analysis, Second ed., Cambridge, NY: Cambridge University Press.
Chen, K. M., Rau, H. H., and Lin, C. C., 2006. The impact of exchange rate movements on foreign direct investment: Market-oriented versus cost-oriented. The Developing Economies, XLIV-3: 269–87
Desai, M. A., Foley, F. C., Forbes, K. J. 2004. Financial Constraints and Growth: Multinational and Local Firm Responses to Currency Crises. NBER Working Paper No. 10545.
Dewenter, K. L., 1995. Do exchange rate changes drive foreign direct investment? Journal of Business, vol. 68, issue. 3. pp. 405– 433.
Fidrmuc, J. and Fidrmuc, Jan. 2000. Integration, Disintegration and Trade in Europe: Evolution of Trade Relations During the 1990s, Working Papers 42,
Oesterreichische Nationalbank (Austrian Central Bank).
Dunning, J.H., 1977. Trade, location of economic activity and the MNE: A search for an eclectic approach. In: Ohlin, B., Hesselborn, P.O. and Wijkman, P.M. Editors, 1977. The international allocation of economic activity. Macmillan, London, pp. 395–418
Dunning, J. H., 2000. The eclectic paradigm as an envelope for economic and
business theories of MNE activity. International Business Review. Volume 9, Issue 2, Pages 163-190
Eiteman, D. K., Stonehill, A. I. and Moffett, M. H., 2004. Multinational Business Finance. 8 edition, Boston: Pearson Addison Wesley. pp 479-492
Froot, K. A. and Stein, J. C. 1991. Exchange rate and foreign direct investment: an imperfect capital market approach. The Quarterly Journal of Economics, Vol. 106, No. 4, p. 1191-1217
Goldberg, L. S. and Kolstad, C. D., 1995, Foreign Direct Investment, Exchange Rate Variability and Demand Uncertainty, International Economic Review, vol.36, issue 4, pp. 855–73.
Harrison, A., 1994. Openness and growth: A time series, cross-country analysis for developing countries. Journal of Development Economies, vol. 48, pp.419– 447. Harris, R.S., and Ravenscraft, D., 1991. The role of acquisitions in foreign direct
investment: evidence from the UD stock market, Journal of Finance. Vol. 46, issue 3: pp.823-44.
Hines, J. R., 1996. Altered States: Taxes and the Location of Foreign Direct
Investment in America, American Economic Review, vol. 86, issue 5, pages 1076-94.
Froot, K. A. and Stein, J. C., 1991. Exchange Rates and Foreign Direct Investment: An Imperfect Capital Markets Approach. The Quarterly Journal of Economics, Vol. 106, No. 4, pp. 1191-1217.
Kiyoshi, S., 1987. The New Manufacturing Challenge: Techniques for Continuous Improvement. New York: Free Press.
Klein, M. W. and Rosengren, E. 1994. The real exchange rate and foreign direct investment in the United States. Relative wealth vs. relative wage effects. Journal of International Economics. Vol. 36. pp. 373-389.
Kohlhagen, S. W., 1977, Exchange rate changes, profitability, and direct foreign investment. Southern Economic Journal, vol. 44, pp.376-383
Kravis, I. B., and Lipsey, R. E., 1982. The location of overseas production and
production for export by US multinational firms. Journal of International Economics, vol. 12, pp.201– 223.
Liu, X., Song, H., Wei, Y. And Romilly, P., 1997. Country characteristics and foreign direct investment in China: a panel data analysis. Weltwirtschaftliches Archiv. Vol.133. pp.313-329
Markuson, J. R., and Svensson, L. E.O., 1985. Trade in Goods and Factor With International Differences in Technology, International Economic Review. Vol.26, No.1.
Mondell, R., 1957. International Trade and Factor Mobility, American Economic Review, XLVII, No. 3, pp. 321-335
Navaretti, G.B., Venables, A.J., 2006. Multinational Firms in the World Economy. Princeton: Princeton University Press.
Pitelis, C., 1996. Effect Demand, Outward Investment and the (Theory of)
Transnational Corporation: An Empirical Investigation, Scottish Journal of Political Economy, Vol. 43, No. 2, pp. 192-206.
Swenson, D., 1993. Foreign mergers and acquisitions in the US. In K.A. Froot (ed.), foreign direct investment, pp.255-86. U.S. Chicago: Chicago University Press. Tinbergen, J., 1962. Shaping the world economy: suggestions for an international
economic policy. New York, NY: Twentieth Century Fund.
Vernon, R., 1966. International investment and international trade in the product cycle. Quarterly Journal of Economics, Vol. 80, pp. 190–201
Wu, Y., 2003. Michael Du Pont. Foreign Direct Investment in Transitional Economies: A Case Study of China and Poland. China Review International, Vol. 10
Xing, Y. 2004. Japanese FDI in China: Trend, Structure, and the Role of Exchange Rates. Working paper. International Development Program, International
University of Japan.
Xing, Y. 2005. Why is China so attractive for FDI? The role of exchange rates. China Economic review. Vol. 17. P198-209.
Xing, Y. and Wan, G., 2006. Exchange Rates and Competition for FDI in Asia. World Economy. Volume 29, Issue 4, P 419 - 434
Yanikkaya, H., 2003, Trade Openness and Economic Growth: A Cross-Country Empirical Investigation, Journal of Development Economics, vol. 72, issue 1, pp: 57-89.
Zhang, J., Jacobs, J. P.A.M and van Witteloostuijn, A. 2007. Multinational enterprises, foreign direct investment and trade in China: the chain of causality in 1980-2003. Journal of Asia Business Studies. Vol. 2, pp.48-57.
Web sites:
American Economic Association (AEA) data base: www.econstats.com (12.10.2008) China Bureau of Statistics :
China Minister of Commerce:
http://www.mofcom.gov.cn/aarticle/s/200612/20061204041019.html (08.07.2008) China Securities Regulatory Committee: http://www.csrc.gov.cn (31.06.2008) China Ministry of commerce, for trade and FDI data: http://english.mofcom.gov.cn/ (01.07.2008)
Hong Kong Economic Information & Agency news agency, for trade and FDI data: http://info.hktdc.com/report/indprof/indprof_060204.htm (13.02.2006)
NCCU Taiwan: http://www3.nccu.edu.tw/~jthuang/FDI.pdf (12.08.2008)
News on Chinese trade and foreign owned enterprise by Xinhua news agency: http://www.chinadaily.com.cn/bizchina/2006-06/25/content_625217.htm (25.06.2006) State administration for industry & commerce of China:
http://wzj.saic.gov.cn/pub/ShowContent.asp?CH=GZDT&ID=334&myRandom=.9779 757533425 (15.08.2008)
Taiwan Stock Exchange Fact Book 2006 :
http://www.twse.com.tw/ch/about/company/download/factbook/2006/web/2.01.htm (09.09.2008)
U.S. bureau for labour statistics, for labour compensation data:
ftp://ftp.bls.gov/pub/special.requests/ForeignLabor/industrynaics.txt (23.02.2007) XinHua Press:
http://news.xinhuanet.com/fortune/2006-08/16/content_4967490.htm (22.06.2008)
