The Empirical Model

In this section we provide an empirical model to estimate the impact of PRC on the inward direct investment of various Asian and Latin American economies. For the East and Southeast Asian empirical studies, we examine Hong Kong, China, Singapore, Taipei,China, the Republic of Korea, Thailand, Malaysia, Philippines and Indonesia. For the Latin American empirical examinations, we include Argentina, Bolivia, Brazil, Chile, Columbia, Costa Rica, Ecuador, El Salvador, Guatemala, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay and Venezuela. The strategy here is to control for all the standard explanatory variables of foreign direct investment in these economies. But we add an additional variable representing the PRC factor. To proxy for the PRC factor, we choose the level of the inflow of PRC's foreign direct investment. Obviously Chinese inward foreign direct investment can also be dependent on the inward direct investment of these Asian and Latin American economies as well as the standard explanatory variables. In order to capture such a reciprocal relationship between the inflow of FDI in PRC and that in other economies, the FDI equation for both sets of these economies and PRC are estimated simultaneously.

The basic regression model for inward foreign direct investment for Asian and Latin American countries and for PRC are written as a linear specification of the following form:

where the subscript "i" and "t" stands for country i at period t and the variables used in this analysis are defined below.

FDI_i,t : the level of inward foreign direct investment in the ith Asian or Latin American economies in year t.
CFDI_t : inward foreign direct investment into PRC in year t.
GGDPR_i,t : growth rate of real GDP of country i at time t.
CGDPR_t : growth rate of real GDP of PRC at time t.
CORRUPT_i,t : an index of corruption of county i at time t.
CCORRUPT_t : an index of corruption of PRC at time t.
DUTY_i,t : import duty of country i at time t.
CDUTY_t : import duty of PRC at time t.
WAGE_i,t : average wage in manufacturing of country i at time t.
CWAGE_t : average wage in manufacturing of PRC at time t.
OPEN_i,t : the share of exports and imports in GDP of country i at time t.
COPEN_t : the share of exports and imports in GDP of PRC at time t.
ILLIT_i,t : the percentage of people who are illiterate of country i at time t.

CILLIT_t : the percentage of people who are illiterate in PRC at time t
TAX_i,t : corporate tax rate of country i at time t.
CTAX_t: corporate tax rate of PRC at time t
GOV_i,t : an index of government stability of country i at time t.
CGOV_t : an index of government stability of PRC at time t.
TEL_i,t : number of telephone mainlines per 1,000 people of country i at time t.
CTEL_t : number of telephone mainlines per 1,000 people of country i at time t
GPCGDP00_i,t : growth rate of per capita GDP (base year 2000) of country i at time t.
CGPCGDP00_t : growth rate of per capita GDP (base year 2000) of PRC at time t
OUTFLOW_t: total outflows of direct investment to the world at time t
LAW_i,t: an index of rule of law of country i at time t
CLAW_t: an index of rule of law of PRC at time t
GDPUSD_i,t: GDP in US dollar in country i at time t
CGDPUSD_t: GDP in US dollar in PRC at time t

The independent variables examined in the analysis are believed to exert an influence on inward foreign direct investment in each country of East and Southeast Asia, Latin America and PRC by changing the investment environment through institutional and policy changes as well as the relevant economic conditions such as the market sizes.

The main variable that we shall examine in this paper is the proxy for the PRC Effect CFDI. There are two sets of arguments that we should consider here. First, in examining which low-wage export platform to locate, multinationals may choose between investing in PRC vs. investing in another country, say Thailand or Mexico. In this case, the multinationals will study the whole host of factors, including wage rates, political risks, infrastructure, etc. that would make a country desirable as a site for lowcost production. Investing in PRC will then reduce the FDI in another Asian or Latin American economy. The sign of CFDI, according to this argument is negative. We shall call this the "investment-diversion effect".

The second aspect is the production and resource linkages between a growing PRC and the rest of Asia and parts of Latin America. In manufacturing, this takes the form of further specialization and growing fragmentation of the production processes. An investor sets up factories in both PRC, Thailand and Mexico to take advantage of their respective competitiveness in distinct stages of productions. Components and parts are then traded among PRC and other economies. An increase in PRC's FDI is then positively related to an increase in Thailand’s or Mexican FDI. Lall and Weiss (2004) document some early signs of an electronics production network between PRC and Mexico.

A different but complementary argument is that as PRC grows, its market size increases and its appetite for minerals and resources also rises. Subsequently, foreign firms rush into PRC to produce in PRC and to sell in PRC. At the same time, other multinationals also invest in other parts of Asia and Latin America to extract minerals and resources to export to a fast-growing PRC in need of a whole spectrum of raw materials. These commodities include copper, steel, aluminum, petroleum, coal and soybeans. This line of reasoning leads one to predict that the sign of CFDI to be positive. We call this effect the "investment-creation effect". Theoretically we cannot determine a prior the net effect of investment-creation and investment-diversion for PRC. It is thus important to examine this issue empirically, as we attempt to do in this paper.

In light of the academic literature that we have surveyed, there are five sets of standard determinants that we will control to isolate the PRC Effect. They are market size variables, labor market conditions, institutional variables, policy variables and the global supply of FDI. These are variables that we identify as important from our literature survey. We will discuss these sets of determinants next.

A substantial literature has developed confirming empirically the importance of the size of the host market and its growth rate. These are measured by GDP, the growth rate of real GDP per capita or real GDP growth. The foreign investors that target the local market are assumed to be more attracted to the country with a higher growth rate of GDP as it indicates a larger potential demand for their products. In the literature, researchers have used both nominal and real GDP measures. As the variables (GDP, the growth of real GDP and per capita real GDP) are used as indicators for the market size and the potential for the products of foreign investors, the expected signs for these variables are positive.

Labor market conditions include the wage rates and the quality of labor. Since the cost of labor is a major component of the cost function, various versions of the wage variables are frequently tested in the literature. A higher wage rate, other things being equal, deters inward foreign direct investment (FDI). This must be particularly so for the firms which engage in labor-intensive production activities. Therefore, conventionally, the expected sign for this variable is negative. However, there are no unanimous empirical results for the effect of labor cost on investment incentives in the existing literature. While some studies have shown no significant role for labor costs, others have shown a positive relationship between labor costs and FDI. The latter result is often attributed to a level of labor productivity or quality of human capital that may be reflected in the wage variables.

The level of human capital is demonstrated to be an another important determinant of the marginal productivity of capital. It has been shown in various studies that skill-related variables are host-country specific. When a host country is more appealing to labor-intensive foreign investment that requires a relatively low level of skills, the importance of the human capital variable tends to be small. On the other hand, labor skills can be a more significant factor for a host country, in which more capital- and technology intensive investment projects are concentrated. In this analysis, we utilize illiteracy rate as a proxy for the level of human capital.

We also examine the significance of institutional factors in the determination of FDI by incorporating the level of corruption, an indicator of the rule of law and an indicator of the stability of each government. Corruption as well as a lack of the rule of law can discourage FDI by inducing a higher cost of doing business. Hines (1995) shows that FDI from the United States grew more rapidly in less corrupt countries than in more corrupt countries after 1977. Wei (1997) presents an alternative explanation of the large negative and significant effect of corruption on FDI. Unlike taxes, corruption is not transparent and involves many factors that are more arbitrary in nature. The agreement between a briber and a corrupt official is hard to enforce and creates more uncertainty over the total questionable payments or the final outcome. Wei demonstrates that this type of uncertainty induced by corruption leads to a reduction in FDI. Political stability of a government and a sound rule of law can also be important factors to foster the inflow of FDI. Uncertain political environments and their related risks can impede FDI inflows in spite of favorable economic conditions. Since the indices of corruption, instability and the rule of law assign higher scores to less corrupt, better law enforcement or a more stable country, the expected signs of the variables, CORRUPT, GOV and LAW, are all positive.

Also included in the analysis are policy-related variables, tariff barriers proxied by import duty, corporate tax rates, openness to foreign trade and the quality of infrastructure. The effect of tariffs on the behavior of multinational enterprises (MNEs) is methodologically demonstrated by Horst (1971). He predicts that in the face of higher tariffs imposed by the host countries, other things being equal, a MNE will increase its production abroad and decrease its exports. More recent models highlight the effect of tariffs on FDI within the context of vertical and horizontal specialization within MNEs. A typical vertical FDI can be characterized by individual affiliates specializing in different stages of production of the output. The semi-finished products in turn are exported to other affiliates for further processing. By fragmenting the production process, parents and affiliates take advantage of factor price differentials across countries. Horizontal specialization on the other hand, involves each affiliate’ engagement in similar types of production. A typical horizontal FDI can be associated with market-seeking behavior and is motivated to avoid trade costs. Choosing between engaging in horizontal FDI or exporting would involve calculating the trade-off between trade costs and economies of scale.

The MNEs, which set up vertical production networks may be encouraged to invest in a country with relatively low tariff barriers due to a lower cost of their imported intermediate products. Therefore, the expected sign of DUTY is negative. In contrast, high tariff barriers induce firms engaging in horizontal FDI to replace exports with production abroad by foreign affiliates (Brainard, 1997; Carr, Markusen, and Maskus, 2001). This "tariff jumping" theory implies a positive relationship between DUTY and FDI. Since the stylized fact about East Asia and Latin America is that a business network is in place in Asia but not in Latin America, the expected sign of DUTY in the Asian regressions is negative, while for Latin America, it is positive (Fukao and Okubo 2003, Ando and Kimura 2003).

OPEN is included to examine the importance of openness of an economy to international trade. The variable measures the degree of general trade restrictions of each country. Following the same line of reasoning above, a negative relationship between openness and market-seeking FDI is expected, and a positive relationship is expected for export-oriented FDI. In addition, in some economies, openness can be an indicator of economic reforms, where domestic reforms and foreign trade reform go hand in hand. FDI can be attracted to a country with more economic reforms.

Another policy-related variable that can influence the host country's location advantage is the host country's corporate or other tax rates. The MNEs, as global profit maximizers, can be assumed to be sensitive to tax factors, since they have a direct effect on their profits. The evidence of significant negative influences of corporate tax rates are reported in previous studies by Wei (1997), Gastanaga, Nugent, and Pashamova (1998), and Hsiao (2001). Better developed regions with a superior quality of infrastructure can also be more attractive to foreign firms relative to others. We test for this by including in our regressions the proxy, the number of telephone mainlines per 1000 people. Fung, Iizaka and Parker (2002) as well as Fung, Iizaka and Siu (2003) show that at least in some instances, FDI is attracted to a Chinese province with a better infrastructure.

Finally, to control for the supply side of the direct investment, we include OUTFLOW, the total global outflows of FDI for each year. An increase in the global supply of FDI can raise FDI in all countries. This can create positive correlations among FDI inflows into various countries that are not related to the PRC Effect. We thus explicitly take this into account. All variables are transformed into logarithms. Data sources and additional explanations of variables are given in Appendix A. The empirical relationship is modeled as a simultaneous equation system and is estimated by the two stage least squares.

Download this Discussion Paper [ PDF 213.2KB| 32 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.