COUNTRY OF ORIGIN DETERMINANTS OF FOREIGN DIRECT INVESTMENTS (FDI) AND FOREIGN TECHNICAL COLLABORATIONS (FTC) IN INDIA

Transcription

1 COUNTRY OF ORIGIN DETERMINANTS OF FOREIGN DIRECT INVESTMENTS (FDI) AND FOREIGN TECHNICAL COLLABORATIONS (FTC) IN INDIA Krishna Chaitanya Vadlamannati * kcv.dcm@gmail.com University of Santiago de Compostela (Spain) ABSTRACT Previous Foreign Direct Investment (FDI) research on India largely focused on examining the determinants of inward FDI. Very limited research has examined the country-of-origin factors of FDI in India. In this paper, we provide the first empirical test of a multidimensional, country-of-origin model of factors related to FDI and Foreign Technical Collaborations (FTC) in India. Economic, Business relations, socio-political and geographic source country characteristics and factor differentials are hypothesized to be important country-of-origin determinants of FDI and FTC in India. Using FDI inflows data for 39 major investing countries in India during the period , the results show that among source country characteristics: economic development, economic reforms, country risk and cost of borrowings are important. With respect to factor differentials with India, factors such as market size, bilateral business relations, skill differences, geographic distance, cost of borrowings and FDI intensity in the past are strongly related to FDI into India. Separate consideration of FTC using pooled binomial regression model for 50 major collaborating countries for the period indicates that technical capabilities, economic development of source countries and bilateral business ties contribute significantly to the explanation of FTCs with India. Keywords: FDI; FTC; India * The DO FILES of the empirical results can be obtained upon request at: kcv.dcm@gmail.com 1

2 1. Introduction Since 1991, India has experienced a tremendous surge in Foreign Direct Investments (FDI henceforth). The total FDI inflows into India increased from 252 US$ millions to over 29 US$ billions in 2007 (UNCTAD, 2008). In fact, India has become an important host country from the standpoint of both FDI inflows and FDI stock. The share of Indian FDI inflows in total FDI inflows of developing countries increased from 0.7% in 1980 to over 25% in 2007 (UNCTAD, 2008). During the early 1990s, India like many other emerging economies replaced its market distorting policies with market creating polices which sought to encourage and attract FDI inflows. Till 1991, FDI into India was restricted and moderately allowed that to only in few sectors. This was mainly because of the kind of policies which the government of India has adopted over the years, includes, `inward looking strategy'; and dependence of external borrowings which inturn resulted in foreign debts, were preferred to the foreign investments to bridge the gap between domestic savings and the amount of investments required. In 1991 when the government of India started the economic reforms program, FDI had suddenly become important for India which was looked upon as a key component of economic reforms package. The New Industrial Policy of 1991 gave utmost priority to attracting FDI inflows. In the process, the government started opening up of domestic sectors to the private and foreign participation which was earlier reserved only for the public sector. This was followed by slow but significant relaxation of regulatory and entry restrictions on FDI inflows. This led to the substantial increase in the volume of FDI inflows into India. During the initial phase of reforms from 1991 to 1998 saw continuous increase in the FDI inflows. The total amount of the FDI inflows during the period to had amounted to US$10,866 million. The increase was largely due to the expanded list of industries or sectors which were opened up for foreign equity participation. However, the inflows declined to the level of US$2,462 million in the year and further to US$2,155 million in The reasons for the declining trend could be attributed to various set of factors, the most important among them being the several restrictions imposed on India by the U.S. on account of the nuclear test carried out by India, the political 2

3 instability, the slow down of the Indian economy due to possible mild recession in U.S. and global economy, the poor domestic industrial environment and the unfavorable external economic factors such as the financial crisis of South-East Asia. From then on FDI inflows surged to over 29 US$ billions as on 2007 (UNCTAD, 2008). There are several studies on India which widely examined the general determinants of raise in FDI and also from sociopolitical perspective. But, studies on country-of-origin determinants of FDI in India are seldom in literature. Past research studies typically laid their focus on India s characteristic features rather than considering both the factors related to source and host country. It is against this backdrop that the present study seeks to explore the links between the conditions of the source countries and their FDI flows into India by examining both source country characteristics and factor differentials across 39 source countries during the period What makes this paper sets apart from other such studies is that firstly, it is first of such kind on India integrating source and host country factors and secondly, along with country-of-origin FDI determinants, we also examine country-of-origin Foreign Technical Collaborations (FTC hereafter) in India using the data for 50 source countries during the period We agree that the rise in FDI inflows and FTCs in India since 1991 is largely attributed to the economic policy reforms initiated by the government of India. However, it is still possible that the macroeconomic and sociopolitical factors in the source country and their differences to those in India can also affect FDI inflows. Hence, we believe that it is not only important to study the general determinants of FDI and FTCs in India, as many of the existing works are already present in literature, but to understand the role played by source country conditions and their relative differences with host country. Thus, we make an attempt in general taking the case of India. The rest of the paper is organized as follows. Section 2 highlights brief survey of literature on similar such studies across the globe. Section 3 illustrates various sets of hypotheses formulated in the study along with information on data sources of all the variables used in empirical models. While section 4 reports empirical results and discussion, section 5 concludes the study. 3

4 2. Literature Survey Attempts to explain the determinants of FDI inflows into a host country were largely based on the motivations and rationale behind firm s overseas expansion and formation. The commonly accepted goal of a multinational corporation (MNC) is to maximize shareholder wealth. Thus, firms enact strategies in order to improve cash flow and enhance shareholder wealth. The company has a target which is an increased market share and, therefore, it proceeds to various local or foreign investments: (a) if the foreign market offers better opportunities (market size, liberalized economy, market prospects, etc.); (b) if the home market (local market, company s market of origin) is too saturated; or (c) due to globalization and competition pressures, in order to increase its sales and profits. Thus, the company and its management must decide when and which market(s) to enter and which entry modes should be implemented. All the possible obstacles that the enterprise will face in the foreign market must be considered and the various incentives offered by the host country (recipient country) must also be taken into account. Since 80% of the worldwide volume of FDI belongs to the multinationals, their attitude may reflect the nature of the world FDI (Bitzenis, 2004). The most important factors which explain the inward of FDI flows to a host country can be broadly classified into micro and macro factors. The micro factors explain the focus of a firm (here it is an MNC) and its motives behind expanding its base and operations and also the means of entering and expansion abroad. The macro factors on the other hand focuses on overall macroeconomic and sociopolitical variables which affect the FDI in general. While the first set of factors deal with firm and industry level studies, the second set of factors examines the country-level factors that determine FDI into a particular country. For example, prominent studies beginning with Aharoni (1966) of U.S. firms abroad; Caves (1974) for Canadian firms and U.K. Manufacturing firms; Globerman (1979) for Canada; Blomstrom (1986) for manufacturing sector in Mexico; Aitken & Harrison (1999) general plat level in Venezuela; and Hsieh (2006) for general plat level for China, have all dealt with micro-level analysis. On the other hand, studies like: Bhagawati, Dinopoulos & Wong (1992); Bhassin et al. (1994); Balasubramaniam et 4

5 al. (1996); Borensztein, E., et al. (1998); Gastanga et al. (1998); Chakrabarti (2001); Baniak et al. (2002); Carstensen and Toubal (2003); Kinoshita and Campos (2003); Janicki & Wunnava (2004); Taahir, R. & Larimo, J (2004); Demekas et al. (2005), explain using macro-level analysis as to why foreign investors prefer investing in one country over another. In further explaining the motives of FDI by MNCs, both micro and macro factors are affected by pull factors and push factors. While pull factors explain why foreign firms prefer one investment destination over the other, the push factors basically explains what forces the firms to expand its operations abroad. Putting things in perspective, pull factors are the host country determinants of FDI, while push factors are the source country determinants of FDI outflows. The pull factors include things like: potential market size, growth of this market size, human capital, absorptive capabilities, exchange rate risk, political and country risk, skills of labour, cost of labour, availability of natural resources, regulations and policy reforms. Some of the push factors are: cost of borrowings, high labour costs, restrictive market environment and so on. Although as highlighted earlier there are numerous studies examining the host country characteristics (pull factors), studies related to source country factors (pull factors) and the differences between the host and source country factors are just handful. A limited number of studies have dealt with determinants of bilateral FDI inflows. First such study is by Ajami & Niv (1984) established the relationship between various national factors influencing the source countries to invest in U.S.A. Their concentration was largely on macroeconomic factors in the source countries. This was followed by Cushman (1985 & 1988) who also included sociopolitical factors along with macroeconomic factors of the source countries. Tallman (1988) then investigated the proposition that home country political risk factors influencing outward FDI of source countries into U.S. The FDI outflows of source countries were regressed on their domestic and international economic and political risk factors. His results show that investment activity from the source countries in U.S. is dependent on home country economic and political conditions. Later on quite a few studies have focused on various set of source country factors influencing their investment decisions in U.S. Prominent amongst them include Froot & Stein (1991); Klein & 5

6 Rosengren (1994); Grosse & Trevino (1996). Their main findings were that the level of economic growth and development, bilateral industry trade, R&D spending, exchange rate issues and cost of capital are the major factors affecting the FDI inflows into U.S. In addition, there are few other studies which have investigated the determinants of FDI on other countries apart from U.S. using source country characteristics. Thomas & Grosse (2001) study the country-of-origin FDI model of economic, socio-political, and geographic factors influencing investments in Mexico. Their results show that factors like: level of bilateral trade, home-country GDP, political risk, geographic distance, and exchange rates, are related to FDI into Mexico. At the same time, relationships between FDI and country-of-origin factors, such as market size (GDP) and cultural distance, that have previously held in research on FDI to large, developed nations do not hold in the same way in the emerging market context. Separate consideration of efficiency-seeking FDI model indicates that bilateral trade, wage rate, GDP, and the exchange rate contribute significantly to the explanation of inward FDI in Mexico. Zhao (2003) taking the case of China conducted a systematic analysis of the connection between FDI and the characteristics of source countries using over 20 years data. The study considers effects of 21 countries factor differentials viz., namely market conditions, risk, and financial factors with China. His results show that, while the market-condition variables and the high values of source country currencies positively influenced the flow of FDI to China, the relatively high costs of capital borrowing and political and operational risks in China inhibited the flow of FDI. Considering the case of 12 industrially developed countries over the period 1981 to 1998, Loungani, Mody & Razin (2002) study the determinants of FDI inflows into 45 developing countries. They conclude that market size of host and source countries, source country specialization, cultural proximity, cost of borrowings drive bilateral FDI. In a very interesting piece of research work by Brito & Sampayo (2005) develop a model that relates the risk, both in home and host countries to the investment decisions. According to their model, higher the risk both at home and abroad deters FDI inflows into host countries. Their results once again reiterates the findings of previous such studies on U.S. 6

7 that countries with higher risk had more difficulties in attracting FDI inflows. A sharp contrast to the earlier findings on bilateral FDI was found by Kimino, Saal & Driffield (2007) for Japan. Using multidimensional model for Japan during , they found statistically weak evidence with regard to the effect of some source country characteristics on FDI in Japan. They find that market size, exchange rates and labour costs on FDI, have a statistically insignificant effect on FDI flows into Japan. They demonstrate that many of the previously theorized links influencing FDI in other developed and developing economies do not necessarily hold and cannot be generalized in the context of Japan. However, they could find that relative exchange rate fluctuation, a higher borrowing cost in investing countries and the stability of the business climate in the investing country as major determinants of FDI in Japan. In more recent studies, Cuyvers, et al. (2008) examined the factors that might affect the inflows of FDI into Cambodia. Panel data sets were used for both approved and realized FDI. The data from seventeen home countries for approved FDI and fifteen home countries for realized FDI were pooled over The home country s GDP growth rate, its bilateral trade with the host country and exchange rate has a positive impact on inward FDI flows into Cambodia. They also find that geographic distance negatively affects the level of FDI inflows in Cambodia. Interesting finding of their study is that the Asian crisis in 1997 and 1998, as well as China s WTO membership are found to have significant negative impact on the Cambodia s ability to attract FDI. Similarly, Zulfiu (2008) study the bilateral FDI flows between Macedonia and rest of the world. His empirical work confirms the expectation of the positive feedback effect of past FDI onto current FDI. Surprisingly, factors like GDP of the host and source country, unit labour cost, trade, inflation, legal environment, distance, and dummy variables capturing the language, common border and colonizing effect, do not have an effect on FDI stocks. 7

8 3. Hypotheses & Data Description The main objective of this paper is to assess the influence of source country characteristic features and factor differentials with host country on FDI outflows to India. For MNCs to operate in an external market, the host country factors like macroeconomic environment, country risk, and cultural distances and so on becomes very important. As MNCs assess the host country factors, it becomes imperative to even take into account the source country factors. In fact the assessment of host country factors would remain incomplete without considering the source country factors. This is because the success of an MNC investments and its operations not only depend on careful assessment and analysis on the host country factors, but also the source country factors because it would act a source of reference to compare the various business, economic and social environmental factors. Thus, the level of host country attractiveness is not just dependent on their factors, but also on the relative differences between the home country and host country factors. As a result, the FDI outflows from the source countries vary with the level and degree of source and host country factor differentials. Following this reasoning, various sets of hypotheses are formulated in this section capturing both source country features and factor differentials between the source and host country. We first formulate hypotheses related to source country factors followed by the factor differentials. FDI inflows into India are largely accounted for by MNCs and other foreign firms especially from developed countries like U.S., Canada and EU countries. Not just in India, but in entire globe, MNCs from these three regions account for higher FDI flows. One important reason for this is that firms from these countries have that kind of massive resources to expand their operations across the globe. They stand to gain from these resources in terms of generating economies of scale. The strategic motives of firms engaging in FDI are three fold. These include, resource seeking, market seeking and efficiency seeking. Thus, large home market enables the firms to expand their base across the globe to accomplish these motives. 8

9 Hypothesis 1: There should be a positive relationship between the market size of source country and FDI inflows into India. Similarly, firms operating in wealthier countries are expected to invest more across the globe. The rationale behind this motive is that firms from wealthier countries are more apparent to internationalize their scale of operations. Also, higher the economic development in the source country, greater is the technological capabilities of their firms. Hypothesis 2: The relationship between domestic wealth of source country and their FDI outflows to India should be positive. Hypothesis 2-a: The relationship between economic development of source country and their FTCs with India should be positive. According to some economists like Tallman (1988) firms operating in a very high risk environment tend to invest more in other countries in order to escape from the risk. This may not be entirely a true argument. Even if it were to be, this argument cannot be generalized to all countries. For example, in the case of India, we consider 39 major investors investing in India. Out of these 39 countries except Thailand, Philippines and Indonesia, compared to rest of 36 countries, India is more risky as an investment destination. This does not prove the point that these 36 countries including U.S., Canada, Australia and all the EU countries that form around 70% of FDI inflows into India are trying to escape the risk in their respective countries. However, what matters certainly is the degree of country risk of India vis-à-vis another emerging market like Mexico, Egypt, Turkey or China. But generally speaking, high country risk certainly deters FDI. India in itself is a best example for this case. During the period of 1989 to 1991, the FDI from major investors who have been investing in India since the independence, the U.S., Germany, U.K. and Japan have all reduced their FDI considerably. This was because India witnessed very high political uncertainty with three governments in just one year, communal tensions, riots and civil wars, stock market scams and more importantly balance of payments crisis which made India virtually default on its external obligations. 9

10 However, other experts like Brito & Sampayao (2005) argue that increase in risk at home country would eventually lead to fewer investments abroad. Using a unique method developed to capture the risk factors both at host and source country, they show that higher the risk on either side or on both sides leads to decline in FDI. Given these divergent arguments, it is not very clear whether we must expect a positive or negative correlation between the source country risk and their propensity to invest in India. Hypothesis 3: The relationship between source country risk and its propensity to invest in India is not clear. Therefore, we assume that there will be a positive / negative relationship between the source country risk and FDI in India. We also capture for a unique variable influencing FDI outflows of a source country, which no other study on multidimensional FDI has captured. We believe that economic policy reforms in the source country are as much an important factor as the policy reforms in the host country. This is because, if the government policies are rigid, marked with higher restrictions, regulations, the cost of doing business for foreign companies in the host country would be higher. From source country s perspective, higher restrictions and regulations in their country would affect productivity of the local firms and human capital as allocation of resources to other sectors becomes restrictive. Likewise, if the investment regulations, which form an important component of policy reforms in every country, are marked with several restrictions and cumbersome rules and regulations would make FDI flows and FTCs discouraging. Indeed this was a strong case in Indian context pre-1990s wherein Indian firms were restricted to raise capital from abroad and the degree of Indian firms operations in global market was also heavily guarded. Policy reforms from 1991 ensured phasing away all the restrictions on local firms to invest abroad. As a result, India as on 2007 is one of the highest FDI investors abroad in Asia. Hypothesis 4: The relationship between economic policy reforms in source country and their FDI outflows to India should be positive. 10

11 Hypothesis 4-a: The relationship between economic policy reforms in source country and their FTCs with India should be positive. Cost of borrowing from the financial institutions and banks constitute a major component of capital cost to the firms. When the cost of borrowings is higher at source country, this becomes an advantage to the host country to attract more FDI provided the cost of borrowings is lower here. Increase in cost of borrowings not only affects the firms operations but would also lead to lose of cost advantages to their competitors in home markets. Hypothesis 5: The relationship between cost of borrowings in source country and their FDI outflows to India should be positive. We now focus on framing hypotheses for the factor differentials between the source countries and India. As highlighted earlier, the size of the market is perceived to be important factor in determining FDI outflows from source country. If the FDI outflows are horizontal in nature, i.e. they are market seeking, then the host country market size also is perceived to be an important factor. MNCs can then enter the national markets having huge market size and may reap the potential benefits. This means that larger the markets, higher the attraction for MNCs which are market seeking. If the host country market size is higher than that of source country, the foreign firms engaged in host country can simultaneously operate in both markets and can also grow concurrently. This is particularly true in the case of India. After China, India has a huge market size amongst emerging markets. From 1994, five years after initiation of market reforms, the market size in India started to grow over 7% on an average per year. Hypothesis 6: The relationship between market size in source and host country together should be positively related to FDI outflows to India. 11

12 Many experts argue that greater the difference between the host country market size and source country, the FDI inflows to the host country would be higher. In fact, Zhao (2003) shows positive relationship between the two. His argument is that when wealth of the source country is comparatively wealthier than the host country, it is likely that there will be more FDI from the source country to explore the business and market opportunities in host country. Going by this argument, we must presume that the host country, especially if it were to be an emerging market represents a potential opportunity to the foreign investors. However, studies like Cravino et al. (2005) argue that after controlling for the joint market sizes of the two, the difference between the market size of source and host countries should be negative. This is because the differences between the two would discourage FDI, if the FDI is a market seeking. Meaning, when one of the countries between the two is smaller, the MNCs would open production facilities mostly in the larger market, thus resulting in negative sign for this variable. Hypothesis 7: Due to the divergent views on market size differentials, we assume that there will be a positive / negative relationship between source and host country market size differences and FDI outflows to India. International trade and FDI are often viewed as complementary to each other. Thomas & Grosse (2001) have shown in the case of Mexico that bilateral trade between the source countries and Mexico lead to higher FDI outflows to Mexico. The argument is that foreign firms when entering into an unknown territory (markets abroad) would try to minimize their risk by resorting to exports to serve the host country market. Later in long run, FDI replaces the exports to serve the market in host country. Many also argue that the relationship between trade and FDI is also bidirectional because MNCs production in host country would give access to untapped markets previously through exporting the goods to those markets. Hypothesis 8: Higher the Exports from source country to India, greater the FDI inflows in India. 12

13 As highlighted earlier the cost of borrowing is the most important component in determining the investments. Lower cost of borrowings gives the firm the edge over both local and firm firms operating in the same market. When the cost of borrowings is higher in source country relative to the host country, firms restrain from borrowing money at home markets (Grosse & Trevino, 1996). Over the last 10 years, various state governments in India are engaged in attracting FDI through their incentives. One of the major components of these incentives includes providing low interest rates on their borrowings within the country. Hypothesis 9: The cost of borrowings in India relative to the cost of borrowings in source country should be positively associated with FDI inflows in India. Another important component which also defines the cost of doing business in India is the labour cost. We could not find the data of labor wages for all the countries in our sample including for India. Therefore, we consider skill differences between the source country and India which also defines the labour costs. The theoretical prediction is that since skill differences should proxy for the differences in labour costs encourages MNCs to integrate vertically in those countries which have very low labour costs. For FDI which is vertical in nature, labour costs play a huge role in attracting FDI like in China. On the other hand, if the skills are higher in the source country relative to the host country, which in a way also defines the wages, there is a strong economic incentive for the firms to engage in operations in that host country. Hypothesis 10: There should be a positive relationship between the FDI inflows to India and the skill differences between the source country and India. But with respect to FTCs it is quite opposite. Higher skill differences means less ability for India to adapt and adopt the foreign technology brought in through FTCs from an advanced country. On the contrary, lower skill differences meaning both countries becoming alike and greater the chances of absorptive capabilities of foreign technologies coming into India. 13

14 Hypothesis 10-a: Higher the skill differences between the source country and India, lower the FTCs with India. However, the empirical evidence on skill differences and FDI is not very supportive. There are studies which could not find any positive association between the two. Moreover, the studies like Blonigen et al. (2003) and Cravino et al. (2005) found negative relationship between the two. In fact Blonigen et al. (2003) highlights the methodological flaw in skill difference variable. They argue that when the skill difference is positive, the source country is more skill abundant than the host country, an increase in this variable means the differences in these countries is rising. On the contrary, if the skill differences are negative, an increase in this variable indicates countries becoming alike. To deal with this issue, we also follow the method of Blonigen et al. (2003) and take the absolute value of the skill differences. We then interact the skill differences with a dummy coding 1 for the years in which the skill differences are negative and 0 otherwise to allow for a different coefficient when the host country (i.e. India) is relatively abundant than the source countries who are investing in India. Hypothesis 11: The negative skill differences interactive dummy between source country and India should be negatively related to the FDI inflows in India. The other most important factor which may explain the variations in FDI inflows into India from different source countries would be cultural differences. The theoretical argument is that the greater the distance between the two cultures, higher the cost of adoption to the local business conditions. The empirical evidence in literature Grosse & Trevino, (1996); Aguiar et al. (2006) also points out negative association on this front. Hypothesis 12: There should be a positive relationship between the FDI inflows in India and cultural proximity between the source country and India. 14

15 Hypothesis 12-a: There should be a positive relationship between FTCs and cultural proximity between the source country and India. Similar to the cultural barriers, another most significant factor is the geographic distances between the source and host country which define the variations in FDI and FTCs. Geographic distances define the transportation, communication and coordination cost, negative sign is on the cards. Studies like Grosse & Trevio (1996); Cravino et al. (2005); Aguiar et al. (2005) have found negative association between the geographic distance between the source countries and host country and the level of FDI inflows. Also, when FDI is market seeking then the distance should have a negative effect on because it involves higher cost of investment and more costly adaptations of goods to local preferences (Johnon, 2006). However, Johnson (2006) argues that distance can be relatively unimportant for resource-seeking investment, since MNCs are attracted to a limited number of geographical locations where the needed resource is available. Hypothesis 13: The relationship between geographical differences between the source country and India should be negatively associated with FDI in India. Hypothesis 13-a: The relationship between geographical differences between the source country and India should be negatively associated with FTCs with India. Business ties and political relations between the two countries also define the level of FDI inflows and FTCs from one country to another. Almost all the studies in the literature on country-of-origin FDI have failed to capture this relationship. We capture this relationship using two variables namely: Bilateral investment treaties and bilateral double taxation treaties between India and source country. We consider not the signing dates, but the dates of ratification which will ensure the enforcement of the treaties. In an interesting study by Haftel (2007) have shown that ratification of bilateral investment treaty between U.S. and other host countries increases the FDI inflows from former to later. On bilateral double taxation treaties, Neumayer (2006) provide evidence that developing countries that have signed double taxation treaties with U.S. or a higher 15

16 number of treaties with important capital exporters actually do receive more FDI from the U.S. and in total. Recently, Barthel, Busse & Neumayer (2008) find that double taxation treaties do lead to higher FDI stocks and that the effects are substantively important. They use a largely unpublished dataset on bilateral FDI stocks of large dyadic panel data. Hypothesis 14: Ratification of bilateral investment treaties signed between India and source country is expected to increase FDI inflows and FTCs to India. Hypothesis 15: Ratification of bilateral double taxation treaties signed between India and source country is expected to increase FDI inflows and FTCs to India. Finally, we believe that the level of investment intensity in the past by the source country in India would have an incentive to invest more in the years to India. This is also for the first time captured here in country-of-origin studies in the literature. By investments intensity we mean that actual strategic investment position of a source country relative to the expected level of investments in India in the future. Methodologically speaking, usage of this variable also helps in controlling for past effects of FDI inflows into India without actually making use of lagged dependent variable 1. Hypothesis 16: The relationship between intensity of FDI between the source country and India in the past should have positive effect on FDI inflows into India. With respect to FTCs, we also make one final hypothesis related to technical capacities of the firms in the source countries. We believe that FTCs with India, which is a developing and progressing economy, can happen on large scale only if the firms collaborating with Indian firms in transferring technology through technical projects enjoy higher levels of technical capabilities back home in their local country. 1 Using a lagged dependent variable, however, may soak up so much variance that theoretically interesting results are likely to get drowned out. Additionally, when a lagged dependent variable is in the model, one is essentially estimating change in the dependent variable from year to year. Such a model complicates the interpretation of results. But many use this method to control for the autocorrelation problem. 16

17 Hypothesis 17: The relationship between technical capacity of the source country and FTCs with India should have positive. To test the hypotheses developed above, two multidimensional regression models are employed viz., Pooled Ordinary Least Squares (POLS) method with two-way random effects for FDI equation function and Negative Binomial (NB) model for FTCs. We choose the negative binomial model over other options for FTCs model for several reasons. First, because our dependent variable is a count, the number of FTCs with India for a given country-year, a linear model is not appropriate. Second, given the proportion of 0 s in our dependent variable and reasonable variance of this variable, the negative binomial is preferred over the poisson, as the negative binomial allows for the possibility of over-dispersion. (for more see: Cameron & Trivedi, 1998). For FDI equation, we use random effects method for OLS because of the possible unobservable effects (Baltagi, 2005). We regard these effects as random and applied two-way random effects estimator which implied that the unobservable effects were part of the disturbance and therefore independent of the observable explanatory variables. Another reason for usage of twoway random effects is because some of the variables like economic reforms which vary at snail pace every year and distance which remains time invariant might be collinear with fixed effects estimations (Beck, 2001). Also, it is not sure whether the unobservable effects are fixed or otherwise. The basic models are: FDI ijt = d 0 + y 1 Source country GDP it + y 2 Country Risk it +y 3 Source country wealth it + y 4 Economic Reforms in source country it + y 5 Cost of borrowings in source country it + y 6 Business relations it + y 7 Exports it + y 8 Market size it + y 9 Difference in Market size it + y 10 Cost of borrowings in source country India it + y 11 Distance it + y 12 Cultural proximity it + y 13 Skill differences between source country & India it + y 14 FDI intensity in past it + ζ it (1) FTC ijt = d 0 + y 1 Source country wealth it + y 2 Economic Reforms in source country it + y 3 Business relations it + y 4 Distance between source country & India it + y 5 Cultural proximity it + y 6 Skill differences between source country & India it + ζ it (2) 17

18 Where, i t = country i at time t ; d = intercept; y = regression coefficients for variable n ; ζ = error term for country i at time t. FDI ijt is the dependent variable in equation 1 measured by the values of log FDI inflows into India from country j in year t in current US$ millions. Two types of FDI values are reported in Department of Industrial Policy & Promotions (DIPP henceforth), government of India: actual FDI inflows and FDI inflows approved. Since FDI approvals are only promises and not original investments, actual FDI inflows are used here. FTC ijt is the dependent variable in equation 2 measured by the count of number of FTCs entered by India with country j in year t. The data is available only in three periods: ; 2003 and The data for FTCs for each country was collected from DIPP from Handbook of Industry Policy and Statistics , published by Government of India. It is noteworthy that the data on FTCs are available only until From thereon, approval of FTCs was merged with automatic FDI route, whose permissions are granted by Reserve Bank of India (RBI) and they do not require any prior permissions. Wealth (Economic development) is measured using log Percapita GDP in current US$ of the source countries. The data comes from World Development Indicators (WDI hereafter), World Bank, Market size; Market size differences and sum of market size measured using the log GDP of the source countries; log GDP of source countries minus India squared and sum of log GDP of Source country and India respectively. Data for GDP is from WDI Economic policy reforms are quantified using Economic Freedom Index constructed by Gwartney, Lawson & Easterly (2006) of Fraser institute. This index is ranked on the scale of 0 (not free) to 10 (totally free). The index captures the most objective measures of liberalization process. This index is comprehensive measure made up of five sub indices capturing: expenditure & tax reforms; property rights & legal reforms; trade reforms; reforms related to access to sound money; labour, business & credit reforms. 18

19 Exports denote the exports of the source country to India. The data on exports is logged and collected from UNCTAD s statistical database on trade & development. Cost of borrowings refers to borrowing costs in the source country, which is the prevailing average lending rate. While the difference between the lending rates in source country and India is the relative difference between the two countries. Average lending rates data was obtained from WDI, Country Risk indicates the operating risk in terms of macroeconomic environment of the country. It is measured using Institutional Investors Magazine s country credit risk index ranked on the scale of While 0 means very high risk environment and 100 is no risk or risk free environment. Bilateral business relations are quantified using the data of bilateral investment treaties and bilateral double taxation treaties between India and source countries. We do not consider the signing date of the treaties but take into account only ratification dates. We account for this using a simple dummy variable wherein 0 is coded if there is no treaty between the two nations. But the value of 1 is given starting from the year when the treaty was ratified. The dates for bilateral investment treaties ratification is taken from UNCTAD data on investment treaties, while the dates for taxation treaty ratification comes from Ministry of Commerce, government of India. Geographic Distance will be measured by the distance from the source country s capital to the capital city of India, New Delhi in miles. The New Delhi is the home to around 15% of the total FDI inflows into India. We used the software developed by Bayers (1999) to estimate these distances. Cultural distance is not measured by an index but, instead, we use dummy variable to measure cultural proximity: language. Because English is the language in which the business operations and government activities are conducted, we code the value of 1 for 19

20 countries whose official language is English and 0 otherwise. The information on official languages of various source countries is collected from CIA World Fact book. Skill Differences refers to the differences of labour skills between source country and India. We proxy secondary school enrollment ratios for skill level in the countries. The data for enrollment ratios are obtained education statistics of UNESCO. The relative differences between the source country and India s enrollment ratios are considered to be the skill differences. FDI intensity in past is used to measure the intensity of the FDI relationship between the source country and India. For this, a ratio is formulated using the trade intensity ratio of Srivastava & Green (1986). The ratio considers the actual value of FDI stock of source country in India with what might be expected in future given the world position of the source country and India. We lag all the values used to formulate the ratio for one year to test whether the past intensity of FDI between the two countries attract more FDI in current year. The formula for this ratio is given as: FDI Intensity Ratio = ActualFDI i j ExpectedFDI i j Where, Actual FDI ij is the FDI inflows from country i to country j and Expected FDI ij is the expected value of FDI inflows from country i to country j. FDI inward stock in country Worldwide FDI inwards j FDI outward stock of country i Worldwide FDI outwards Worldwide FDI inward Thus, if the intensity ratio is greater than the value of 1 it means that the FDI intensity of country j in country i is very high. Meaning, the FDI relationship between the two countries is stronger than expected. If the ratio is less than 1, it means the opposite. 20

21 The pooled time-series cross-sectional (TSCS) data may exhibit Heteroskedasticity and serial-correlation problems. But these problems do not bias the estimated coefficients as pooled regression analysis in itself is a more robust method for large sample consisting of cross section and time series data. However, they often tend to cause biased standard errors for coefficients, producing invalid statistical inferences. To deal with these problems, Beck and Katz (1995) propose to retain POLS parameter estimates but replace the POLS standard errors with panel-corrected standard errors (PCSE). They find that these estimates of sampling variability are very accurate, even in the presence of complicated panel error structures. Following others, this analysis employs POLS regression with PCSE Cross-section weights. 4. Empirical Results & Discussion The sample of country-years that we examine in total is 592 observations 2. The results of regression estimates using random effects method in assessing the country-of-origin determinants of FDI in India are presented in 6 different models in table 1. We also control for heteroskedasticity using Cross-section weights PCSE heteroskedasticityconsistent standard errors & covariance. The summary of data is provided in annexure 3. The model 1 in table 1 deals with only source country characteristic features and does not include the factor differentials between India and various source countries. The results are in accordance to the theoretical predictions made in the hypothesis section. Improvement in country risk in source country is negatively associated with inflows of FDI into India. Though we could not find any statistical significance for this variable in model 1, we find that it is consistently significant in rest of the models (see model 2 to model 6). The country risk scores suggest that on a scale of 0 to 100, zero represent high country risk and 100 represent low or no country risk. Therefore the negative effect of country risk suggests an improvement in country risk of source country. Our findings on the negative effect of source country risk on FDI inflows into India are in line with other studies like: Tallman (1988) and Grosse & Trevino (1996) who advocate that countries with higher country risk (political and economic) deter investments into a host country 2 Due to unbalanced data, the total number of observations has come down from 624 to

22 from abroad. These results with respect to country risk strongly support the view that firms operating in countries with higher internal risk would try internalize seeking to escape from the home country risk. The results also demonstrate that higher economic reforms in the source country are positively associated with FDI inflows into India. An increase in economic reforms index by 1% points in source country leads to increase in FDI inflows by 0.86%. This means that improvement in policy reforms in source country explains around 80% variation in their investment flows to India. These results pertaining to reforms are so strong across the board that irrespective of introducing various control variables, the coefficient value does not change and so do the statistical significance which is at 1% confidence level. Table 1: Determinants of FDI in India - country by origin Dependent variable: Log (FDI inflows) from Source country to India Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Constant * (4.97) * (5.17) * (5.62) * (5.69) * (7.989) * (4.39) Source Country Risk (0.01) *** (0.01) (0.01) *** (0.01) ** (0.072) * (0.01) Economic Liberalization in Source Country * (0.30) * (0.30) * (0.31) * (0.29) (0.688) * (0.21) Economic Liberalization Country Risk ** (0.010) Log (GDP) of Source country * (0.21) Log (GDP of Source country + India) * (0.41) * (0.43) * (0.41) * (0.423) * (0.30) Log (GDP differences) between Source country & India (0.08) (0.09) (0.08) (0.086) (0.06) Log (Percapita GDP) of Source country * (0.31) *** (0.32) (0.34) (0.32) (0.333) * (0.18) Bilateral Investment Treaties ratified * (0.24) * (0.23) * (0.23) * (0.23) * (0.242) * (0.22) Bilateral Double Taxation Treaties ratified *** (0.34) (0.35) ** (0.37) (0.34) ** (0.371) * (0.27) Log (Exports from Source country to India) (0.11) (0.11) (0.11) (0.11) (0.117) (0.07) ***

23 Cost of Borrowing in Source country (0.00) Cost of Borrowing between Source country & India *** (0.00) *** (0.00) *** (0.00) (0.005) * (0.00) Cultural Proximity between Source country & India *** (0.50) (0.49) (0.53) (0.49) (0.524) (0.24) Log (Distance in miles) between Source country & India (0.72) ** (0.66) ** (0.70) ** (0.71) ** (0.703) (0.46) FDI Intensity in past * (0.18) * (0.17) * (0.16) * (0.18) * (0.177) ** (0.58) Skill Differences between Source country & India * (0.00) * (0.00) * (0.00) * (0.007) (0.00) Skill Differences Negative dummy of skill differences * (0.09) Log (total FDI Outflows from Source country) ** (0.10) Europe Dummy * (0.15) Asia Dummy ** (0.61) R-squared Adjusted R-squared F-statistic * * * * * * Number of Countries under study Total Number of Observations Country Dummies YES YES YES YES YES YES Period Dummies YES YES YES YES YES YES Note: * Significant at 1% confidence level; ** Significant at 5% confidence level; *** Significant at 10% confidence level; + Significant at 15% confidence level. All models are controlled for Heteroskedasticity. Cross-section weights PCSE Heteroskedasticity-consistent Standard Errors are reported in parenthesis. In fact, in of the rare studies on determinants of FDI outflows, Kyrkilis & Pantelidi (2003) shows that higher the trade and investment openness of the source country, greater the FDI outflows. Our study supports their finding by introducing an appropriate proxy for openness of the source country, which is economic policy reforms process. The results also highlight significant positive impact of source country s economic growth on FDI outflows. An increase in log GDP levels by 1% lead to 0.75% increase in FDI outflows to India. This is in line with the theory which shows that a very high level of FDI flows is accounted by large firms that have the resources to expand their operations abroad and stand to gain from economies of scale. Indeed world s largest MNCs come from countries like U.S.; Japan; U.K; Germany and France, which have very high home country market size. Since more than 80% of FDI inflows into India come from U.S., 23