ASSESSING THE RELATIONSHIP BETWEEN TRADE AGREEMENTS AND FOREIGN DIRECT INVESTMENT

ASSESSING THE RELATIONSHIP BETWEEN TRADE AGREEMENTS AND FOREIGN DIRECT INVESTMENT A Thesis submitted to the Faculty of the Graduate School of Arts & Sciences at Georgetown University in partial fulfillment of the requirements for the degree of Master of Public Policy in the Georgetown Public Policy Institute By Samuel Ross Easterly, B.S. Washington, DC April 7, 2009

ASSESSING THE RELATIONSHIP BETWEEN TRADE AGREEMENTS AND FOREIGN DIRECT INVESTMENT Samuel Ross Easterly, B.S. Thesis Advisor: Michael Clemens ABSTRACT This paper analyzes the relationship between preferential trade agreements (PTAs) signed between countries and the foreign direct investment (FDI) inflows to the member countries of the agreements. Using the most comprehensive database of PTAs available, it extends earlier research by considering the relationship between FDI flows and different types of bilateral and multilateral trade agreements, by controlling more carefully for political institutions, and by analyzing FDI flows between pairs of countries rather than FDI receipts. FDI has grown in importance to economies across the globe and establishing a link between PTAs and FDI would provide policymakers with another avenue to promote FDI inflows to their countries. This study demonstrates that institutional variables play a role in the relationship between trade agreements and FDI, and that relationship also differs depending on the type of trade agreement for example, entering a customs union with an OECD country has a very different relationship with FDI than joining the WTO. ii

I would like to thank my thesis advisor, Michael Clemens, for his invaluable support, expertise, and guidance throughout the entire process, without which this thesis would never have been completed. I would also like to thank my former colleagues at the U.S. International Trade Commission for their generous assistance in helping me find a thesis topic. I would like to thank all of my professors from the Georgetown Public Policy Institute who taught me the concepts and skills necessary to complete this type of analysis. Last, but certainly not least, I would like to thank my friends and family who have supported me throughout my education at the Georgetown Public Policy Institute and while completing the thesis-writing process. iii

TABLE OF CONTENTS Introduction...1 Background...8 Literature Review...12 Market Size and Characteristics...12 Macroeconomic Conditions...14 Political and Institutional Variables...15 Population and Environmental Factors...16 The Theory of Preferential Trade Agreements...18 Trade Agreements and FDI...19 Conceptual Model...24 PTAs Could Cause FDI...24 Conditions that Could Cause both PTAs and FDI...25 Conditions that Could Be Caused by FDI and PTAs...28 FDI Could Cause PTAs...31 Analysis Plan...32 Data Description...36 Descriptive Statistics...38 Regression Results...41 Results...41 Aggregate FDI Flows Regressions...41 Bilateral FDI Inflows from All Countries (Senders) to OECD Countries (Receivers)...46 iv

Bilateral FDI Outflows from OECD Countries (Senders) to All Countries (Receivers)...53 Economic Interpretation and Policy Implications...56 Total FDI Flows Regressions...56 Bilateral FDI Inflows from All Countries (Senders) to OECD Countries (Receivers)...61 Bilateral FDI Outflows from OECD Countries (Senders) to All Countries (Receivers)...70 Conclusions...74 References...78 v

LIST OF TABLES Table 1 Model Variables...32 Table 2 Aggregate FDI Flows Model Descriptive Statistics...38 Table 3 Bilateral FDI Flows Models Descriptive Statistics...39 Table 4 Aggregate FDI Flows Regressions...44 Table 5 Bilateral FDI Inflows from All Countries to OECD Countries Regressions...47 Table 6 Bilateral FDI Outflows from OECD Countries to All Countries Regressions.54 vi

Introduction In recent years, foreign direct investment 1 (FDI) has acquired increasing importance to economic strength in both investor and recipient countries. After four straight years of growth, in 2007, global FDI inflows increased by 30% to $1.8 trillion, a record high (UNCTAD, 2008). Nations view FDI as a beneficial source of foreign capital in part because it is associated with a transfer of foreign technology and skills in areas like managerial techniques, marketing ideas, accounting practices, and multiple other business-relevant realms. The near consensus in the FDI literature is that, except in cases of serious market distortions, FDI increases income and social welfare in the host country (Moosa, 2002). Developed countries are the main hosts for flows of FDI, at around sixty-eight percent of the global total in 2007, and their economies benefit accordingly. Jackson (2008), for example, argues that FDI contributes to job creation, wage increases, growth in the manufacturing sector, access to new technologies and skills, increases in labor productivity, more tax revenue, and lower interest rates. Yet, at the same time, FDI flows also benefit developing countries. Flows to developing countries reached a record high in 2007 at around $500 billion, and developing countries continued to grow as an important 1 The definition of foreign direct investment taken from UNCTAD s 2008 World Investment Report: Foreign direct investment (FDI) is defined as an investment involving a long-term relationship and reflecting a lasting interest and control by a resident entity in one economy (foreign direct investor or parent enterprise) in an enterprise resident in an economy other than that of the foreign direct investor (FDI enterprise or affiliate enterprise or foreign affiliate). 1

source of FDI flows, which reached a record level of $253 billion in 2007 (UNCTAD, 2008). The benefits of FDI manifest themselves a little differently in developing countries than they do in the developed world: FDI is more reliable than equity and debt capital flows, providing needed stability to developing economies; FDI has a positive relationship with savings and investment without deterioration in the current account; FDI and GDP growth are significantly correlated and develop further over time; and FDI can boost trade for developing countries through export promotion (foreign investors build plants where their goods can be produced at lower costs) and diversify exports when developing countries use technology from developed countries to produce goods normally sold by the developed world (Kumar, 2007). The nature of FDI poses a distinct challenge to the actual investors, mainly multinational companies (MNCs). FDI often entails a physical resource built by an MNC in the foreign/host country, which involves reduced liquidity and its associated higher risk. This higher risk applies to economic concerns with the investment, such as fluctuations in the exchange rate or interest rate, and political concerns, such as the possibility of expropriation or regime change that creates an inhospitable business climate. Investors will naturally want to find locations with the lowest possible risk, all other factors held equal, and nations interested in attracting FDI will present their investment environment as financially, politically, and socially stable and economically prosperous (Hicks, 2007). Over 500 investment promotion agencies in over 160 2

countries are responsible for disseminating information about investment opportunities in the country, improving the overall investment climate, and creating a positive image of the country abroad (Zanatta et al., 2006). Additionally, national governments can provide tax breaks and subsidies, and allow foreigners majority ownership of projects, free repatriation of profits, liberal employment of expatriates, and any number of other incentives to allay investor concerns related to their potential investment. Another way which will be the focus of this paper in which a nation can prove to the global investment community that it is financially secure and politically stable is its membership in a preferential trade agreement (PTA). Between 1989 and 2008, the number of PTAs notified to the World Trade Organization (WTO) and in force in 2008 increased by 385% from 46 to 223 agreements. The PTAs signed over the past two decades significantly differ in content from older agreements in that many of them are not exclusively concerned with merchandise trade, but instead try to integrate other areas like investment, trade in services, setting and harmonization of standards, competition disciplines, customs cooperation, intellectual property rights (IPR), and dispute settlement (Medvedev, 2006). The United States, for example, signed bilateral trade agreements this decade with Jordan (2000) and Morocco (2004). Neither agreement was critical to the United States from an economic perspective. In the year before each agreement, Jordan was the 100 th largest trading partner (in terms of the sum of imports from and exports to Jordan) and Morocco was the 80 th largest trading partner. However, both FTAs had important geopolitical implications 3

in further cementing U.S. ties in the Middle East and North Africa. Both agreements also included many deep integration issues such as an IPR chapter in the Jordanian FTA and an investment chapter in the Moroccan FTA. The growing prevalence of PTAs and, more importantly, PTAs that go beyond lowering barriers to merchandise trade, opens up a great number of research questions with important policy implications that can be empirically studied. This paper investigates whether or not analysis of data from the past two-and-a-half decades on trade agreements and FDI flows can reject the following null hypothesis: The effect of preferential trade agreements on foreign direct investment flows is not statistically significantly different from zero. This hypothesis has major policy ramifications for trade representatives and politicians in developed and developing countries who are negotiating trade agreements. While the results may or may not have external validity and instead be country-specific, in general, if the effect of PTAs on FDI inflows is found to be statistically significantly different from zero and positive, developed countries will have found a potential avenue to boost their economies outside of the traditional fiscal and monetary measures, while developing countries will have found a source of stability, technology and information-sharing, and economic growth. This paper builds upon and improves on the existing literature in several important ways. First, this paper not only studies PTAs, but also WTO accessions. Becoming a member of the WTO may be a greater sign of policy reform and stability than implementing a PTA, and, therefore, a better signal for investors to send FDI to that 4

country. Secondly, this paper considers the relationship between institutional variables and FDI flows, such as the level of human capital and level of political constraints in government. A country with high levels of human capital or a constrained system of governance with checks and balances might be more likely to attract FDI inflows. Thirdly, this paper attempts to determine whether or not different types of trade agreements have different associations with FDI. Trade agreements with the highest levels of integration (e.g.., customs unions and common markets) might also attract higher levels of investment. Finally, this paper examines the relationships associated with bilateral FDI flows specifically to and originating from the member countries involved. Past work has examined the association between trade agreements and overall FDI flows, but it would also be useful to determine whether most of the association between trade agreements and FDI flows could be attributed to the member countries of the trade agreement, since investors in those countries would have the most intimate knowledge of the investment climate and the effect of the trade agreement on it. This paper finds that membership in the WTO has a bigger partial relationship with FDI flows than the other trade agreement variables, and that in the bilateral FDI models, joining or membership in the WTO is associated with a decline in FDI flows from any one particular country because the country has multiple destinations and sources for investment. From a policy perspective, countries should spend more time negotiating agreements under the WTO than pursuing bilateral or regional agreements in their country s trade policy, for example, by reviving the Doha round of negotiations. 5

This paper also finds that human capital and political constraints have a significant correlation with FDI flows. The results indicate that as the primary enrollment ratio in a country increases, it is associated with attracting higher levels of FDI and sending lower levels of FDI, as the economy of the country improves to the point where it can improve by itself with less foreign investment. The results for political constraints indicate that a more constrained system of government is associated with higher FDI flows, but only to a certain point that most developed countries have likely already met. Consequently, countries should continue efforts to improve their levels of education and, over the long run, countries should work towards a system of government with an adequate level of checks and balances. Finally, this paper finds that involvement in a trade agreement does have a positive partial relationship with FDI flows. However, the results are more nuanced when broken down by trade agreement type. For example, in the aggregate FDI model, only customs unions have a statistically and economically significant relationship with FDI. Similarly, in the bilateral FDI models, only trade agreements classified as Other and free trade agreements have statistically and economically significant relationships with FDI, with the former having a larger partial relationship when statistically significant in the OECD country FDI-sending model and specifications. Consequently, if countries choose to pursue bilateral or regional trade agreements outside of the WTO, they should try to make them as economically integrative as possible within the normal limits of political feasibility in order to attract higher levels of aggregate FDI irrespective 6

of the source. However, countries can choose to pursue much more basic agreements when considering bilateral trade strategies. 7

Background Trade agreements over the past fifty years can be classified into three waves. The first wave of agreements from the 1950s to late 1960s was fairly limited in scope and, since tariff levels were higher in this period than later, the preferential liberalization of merchandise trade was the main goal. The first wave included the first attempts by developing countries to create PTAs to help them lower the costs of their import substitution industrialization policies (Adams et. al., 2003). The second wave began in the 1980s as the United States entered into several PTAs. Previously, the United States had opposed to PTAs due to its support of the General Agreement on Trade and Tariffs (GATT) principle of Most Favored Nation (MFN), which holds that any trade concession granted to individual members should be extended to all members of GATT. Agreements in the second wave of trade liberalization included some non-traditional areas, such as dispute resolution, but the main focus remained on lowering tariffs and establishing free trade areas (Adams et. al., 2003). The third wave of trade agreements began in the 1990s with an explosion of PTAs worldwide. The key difference between third-wave and earlier agreements was that the majority of them began to include non-traditional areas such as investment, trade in services, setting and harmonization of standards, competition disciplines, customs cooperation, intellectual property rights (IPR), and dispute settlement (Medvedev, 2006). In terms of investment, provisions such as dispute settlement mechanisms and protection of intellectual property rights send a clear signal to foreign investors of the 8

institutional stability associated with the agreement and nations involved. While none of the provisions is necessary or sufficient to increase foreign direct investment (FDI), observers suggest that together they are likely to create investment (Medvedev, 2006). Indeed, when an agreement does not include any of the non-traditional areas, it could still provide a positive signal to investors by solidifying existing reforms. From this perspective, PTAs could be viewed as a stepping stone towards global free trade (Krueger, 1999) and evidence of participating countries greater outward orientation, leading investors to believe the countries will have lower political risk, and boosting FDI (Medvedev, 2006). One of the most important benefits of FDI to developed economies and, over the past two decades, to developing economies as well is the role it can play in transferring technology between nations. This boost to technology could help to foster research and development, create economies of scale, and lead to technological spillover effects. Technological spillover effects, for example, can work through several channels: 1) local firms adopt technologies introduced by multinational firms through reverse engineering or imitation; 2) workers trained by the multinational may transfer knowledge by switching employers or creating their own firms; and 3) multinationals transfer technology to firms that act as suppliers of intermediate goods or buyers for their own products (Saggi, 2000). FDI can benefit economies through more traditional economic means as well, for example, by creating new jobs and boosting growth in GDP, and, especially for developing countries, providing a source of illiquid investment that, once 9

invested, assures the host country of its long-term existence and associated increase in economic stability (Ahlquist, 2006). Policies that attract FDI flows are particularly attractive to government officials in both developed and developing economies. In the past decade, government officials interested in pursuing FDI promotion policies tried to use the WTO as a forum, especially in the Trade and Investment negotiations in the Doha Round. However, this most recent round of multilateral negotiations at the WTO broke down in late July 2008 and is not expected to resume until some point in 2009 at the earliest. The Doha round was dubbed the development round for its intended focus on issues most pressing to developing nations. It also had tried to focus on deeper integration issues for developed countries. After the failed Doha ministerial conference in August 2004 in Cancun, however, most of these issues were dropped from negotiations and the world lost an opportunity to balance the interests of countries where foreign investment originates and where it is invested, countries right to regulate investment, development, public interest and individual countries specific circumstances. (WTO, 2008) In terms of FDI policies at the country level, protectionism has slowly been rising in recent years. Several countries have tightened investment rules or enacted new rules to regulate FDI and protect strategic sectors. Host-country protectionism has led to an increase in skepticism towards and regulations for cross-border mergers and acquisitions. And even home-country protectionism negative attitudes towards outward flows of FDI has risen, as outsourcing has become the prime example of assumed ills associated 10

with globalization (Economist Intelligence Unit, 2008). In light of these developments, if analysis can demonstrate a link between PTAs and increased FDI flows to the member countries, policymakers in developed and developing countries still have an avenue to pursue an FDI promotion policy. This policy could circumvent the static WTO forum, while bringing some of the more traditional benefits of trade liberalization through bilateral or pluri-lateral trade agreements. Most importantly, the countries would be sending a signal to investors that, even in a time where protectionist pressures against FDI have slowly risen, their countries still provide a safe environment in which to invest. 11

Literature Review The extensive literature on the many potential determinants of FDI flowing into a country can be broken down into the four categories of market size and characteristics, macroeconomic conditions, political and institutional variables, and population and environmental factors. Dunning (1977) contributes a model for FDI determinants that has been cited often in the literature: his OLI (ownership-location-internalization) model. According to Dunning, FDI occurs when three conditions are met: 1) a multinational company (MNC) has an ownership advantage through which it is competitive in that market where it wants to invest; 2) one place has to have a location advantage (e.g., a large domestic market for the investing company s goods) over another for an MNC to invest there; and 3) there has to be an internalization advantage where owning a plant in another country is better than licensing agreements with a firm based there (Kolstad and Tøndel, 2002). Market Size and Characteristics The literature on FDI determinants supports a strong relationship between the size of the host country market and the amount of FDI that market receives. According to the UNCTAD (1998), larger markets are better able to accommodate increased investment, both domestic and foreign, because they have large numbers of firms, and because firms have more opportunities to develop scope and scale economies. In a literature review of surveys and econometric studies, Lim (2001) notes that the most robust determinant [of FDI] is the size of the market. Market size proxied by real GDP is highly significant 12

and positive in virtually all the studies. Lim further describes the effect of market size in terms of the intended effect of FDI by showing that larger host markets will likely attract horizontal FDI (entire production process in a new country), but they will be indifferent to vertical FDI (part of the production process in a new country). UNCTAD (1998) proposes a slightly different explanation of FDI based on the motives of the company. Still, one of the economic determinants is market-seeking (horizontal FDI that includes market size and per capita income in the list). Blömstrom and Kokko (1997) believe that larger market size allows firms to grow larger than they would have in purely national markets, which could allow the firms to invest in more R&D and marketing, which may lead to the creation of new intangible assets that stimulate FDI, within as well as outside their own region. Another market characteristic in the FDI determinants literature is the geographic distance between the country markets. Models of bilateral FDI flows, such as the one used by Adams et al. (2003), find a significant negative impact of distance between recipient and sending countries on FDI stocks. However, Medvedev (2006) postulates that if trade and FDI are treated as substitutes by firms, there could be a positive relationship between distance and FDI. Blonigen s (2005) literature review on FDI determinants shows that the gravity specification traditionally used to predict trade flows between countries as primarily a function of the GDP of each country and the distance between the two countries also fits patterns of FDI reasonably well. However, there is no support for a model of FDI flows with gravity variables as the sole determinants since 13

intuition and theory suggests that FDI behavior is likely much more complicated to model than trade flows. Macroeconomic Conditions The literature on FDI determinants is heavily focused on macroeconomic-related factors. The relationship between the level of trade openness in a country and its association with FDI, for example, has been examined by many authors. According to Blonigen (2005), higher trade protection should give firms an incentive to relocate production to the country to which they were exporting in order to avoid the higher costs from the trade protection. This phenomenon, termed tariff-jumping, has been documented empirically (Blonigen, 2002), and applies directly to horizontal FDI. Conversely, lower trade protection may benefit vertical FDI because the easier it is to enter the country the more a firm can take advantage of lower transportation costs and cheaper resources (Hicks, 2007). In addition to this difference related to openness, FDI and trade protection may be endogenous because a nation s policies for trade protection could explicitly target certain import sectors where FDI is less likely (Blonigen, 2005). Stability in a country s exchange rate is also a macroeconomic determinant of FDI. UNCTAD (1998) states that [e]xchange-rate policy is related to stability and may influence FDI decisions by affecting the prices of host country assets, the value of transferred profits, and the competitiveness of foreign affiliate exports. In addition, Froot and Stein (1991), using annual U.S. aggregate FDI data in an empirical study, show that currency depreciation is associated with higher inward FDI. This depreciation will 14

increase the relative wealth of foreigners and increase the relative rate of return for foreign firms investing in domestic assets since they can avoid paying a domestic monitoring penalty, thereby further encouraging additional FDI (Medvedev, 2006). Finally, Blonigen (1997) also affirms the positive link between foreign exchange rate depreciation and FDI, reasoning that if foreign firms purchase another country s assets, they can generate returns from those assets to their benefit in currencies other than those used for the purchase. Political and Institutional Variables Quéré et al. (2005) are exclusively concerned with institutional determinants of FDI. They find that institutions matter independently of GDP level and that public efficiency is a major determinant of FDI. The authors suggest three general reasons why institutions could matter for attracting FDI: 1) good governance raises productivity prospects, which is attractive to investors; 2) poor institutions are associated with problems such as corruption that would bring extra costs for investors; and 3) since there are high sunk costs involved with FDI, any form of uncertainty stemming from poor government efficiency, policy reversals, graft or weak enforcement of property rights and of the legal system in general is particularly damaging to FDI. The authors construct a gravity model with bilateral stocks of FDI as the dependent variable, and find that institutions such as tax systems, ease of creating companies, level of corruption, transparency, contract law, security of property rights, and efficiency of justice and prudential standards are all key determinants for FDI. Other studies, including one by 15

Globerman and Shapiro (2002), come to the same conclusions about the importance of institutional variables for FDI flows. All of these studies, however, note the important caveat that data for institutional variables is notoriously difficult to capture. A growing literature also focuses on the impact on FDI of countries joining institutions, such as PTAs. Büthe and Milner (2005), for example, study the impact of signing bilateral investment treaties, joining the GATT/WTO, and joining a PTA. They claim that joining these treaties and organizations requires countries to undergo economically liberalizing reforms, as well as making it much less likely that a country will renege on an economic commitment since there are now enforceable rules and another country or countries (in the case of the WTO, potentially over one hundred countries) can punish the offending country (Hicks, 2007). Population and Environmental Factors Population and environmental factors may also play a role in determining the level of a country s FDI. Kolstad and Tøndel (2002) note, for example, that MNCs could be attracted to areas with low levels of social development and equality if the labor is cheap. They could be attracted to areas with high levels of human capital accumulation if this condition is associated with higher levels of productivity. Therefore, proxy variables to account for the level of social development of a country s population, such as the percentage of people who have completed secondary education, are frequently used in the FDI-determinants literature (Kolstad and Tøndel, 2002; Globerman and Shapiro, 2002), 16

and research indicates that their impact depends on the industry in which the investment is made. Aside from the social or human development characteristics of a country s population, other factors that are almost entirely country-specific that can be associated with FDI flows. For instance, historically, the most important determinant of FDI was the level of a country s natural resources. Although this factor has declined in significance as the importance of the primary sector in world output has declined, it can still explain a significant portion of inward FDI in natural resource-rich developing countries (World Investment Report, 1998). In addition, Adams et al. (2003) find that a number of country-specific variables are significant in their bilateral FDI model, such as how similar the two countries languages are (positive correlation with FDI), whether they have colonial ties (positive), whether they share a border (positive), and whether one or more of the countries is landlocked (negative). Investors also perceive potential threats to their investment from the political environment and act accordingly. As intrastate conflict rises, whether in the form of political turmoil or, more significantly, actual combat, the level of FDI flows to the host country drops (Nigh, 1986). Conversely, some evidence indicates that if the host country is a developed nation military conflicts fought on foreign soil might encourage FDI since they have the financial capacity to fight the war and potentially need FDI to help fund it (Hicks, 2007). 17

The Theory of Preferential Trade Agreements The three waves of PTAs have each given rise to a distinct body of theoretical work. Theoretical work during the first wave focused on the static effects of trade agreements and whether or not they would bring benefits to individual countries, the group as a whole, and/or to countries left out of the agreement. Research during the second wave tried to determine whether PTAs were building blocks or stumbling blocks to the development of the multilateral trade system. Finally, in response to the third wave, the focus has started to shift to effects of the non-trade provisions of PTAs. The following discussion of these three bodies of work draws exclusively from Adams et al. (2003). The most straightforward theoretical explanation of PTAs concerns trade creation vs. trade distortion effects. PTAs reduce the average tariff on imports for goods entering a country, thereby lowering one source of economic distortion. Yet, they simultaneously increase another distortion by making a country s tariff schedule less geographically uniform. Consequently, PTAs can improve a country s economic welfare by shifting production from a higher-cost domestic source to a lower-cost PTA partner trade creation. However, they can also reduce a country s economic welfare by shifting production from a low-cost non-member to a higher-cost PTA partner trade diversion (Adams et al., 2003). Much of the literature devoted to the first wave of trade agreements tried to establish general rules of thumb that would exist in situations where the gains from trade creation in PTAs would exceed the losses from trade diversion. Unfortunately, 18

this analysis has proved generally fruitless as governments cannot easily identify PTA opportunities that meet this criterion. Four studies examine incentives for PTAs to expand their membership and whether or not that expansion will be a stumbling block or a building block to multilateral liberalization. Most literature indicates that PTAs are stumbling blocks to multilateral trade liberalization. Krishna (1998), for instance, concludes that PTAs will be stumbling blocks to the multilateral negotiations because they reduce incentives of members to liberalize tariffs with non-member countries. Levy (1997) also finds that bilateral PTAs can undermine political support for multilateral free trade, and when multilateral liberalization is not feasible without a PTA, it almost certainly will not become feasible in the presence of a PTA. Baldwin (1996) argues that PTAs serve as a building block because firms in non-member countries will see their profits decline, and will then lobby their government to enter the PTA, pushing the balance towards entry in the countries at the margin of that decision process. However, Zissimos and Vines (2000) argue that member countries have an incentive to prevent new entrants into the PTA in such a fashion that PTA formations will still fall short of multilateral liberalization. Trade Agreements and FDI Over the past ten years, a small but growing number of studies have focused on the effects of PTAs in shaping FDI flows. This literature is the starting point for the present paper. 19

Blömstrom and Kokko (1997) examine the investment effects of regional integration in the case of three regional trade agreements: the Canada-U.S. FTA, NAFTA, and MERCOSUR. The authors hypothesize that the FDI process for a country in the context of regional integration can be mapped onto a basic template with the level of environmental change (degree to which trade and investment flows are liberalized by the agreement) as weak or strong and the advantage of location (the degree to which it is more profitable to locate a firm s economic activity in a location) as weak or strong. In this observational study, the authors do not run regressions, but conclude that the evidence from the three agreements supports their rough hypothesis that an agreement in a strong position of both environmental change and advantage of location is more likely to lead to inflows of FDI from countries both outside and within the agreement. Later studies have moved beyond theoretical and observational examinations to empirical research. One of the most important contributions to this line of research is that of Adams et al. (2003) who examine the trade and investment effects of PTAs. These authors use a gravity model in which the dependent variable is the natural logarithm of the stock of outward investment from home country to host country in a number of developed and developing countries from 1988 to 1997. Their results show that six of the nine PTAs were investment-creating, one was investment-diverting, and two showed no clear impact. They also conclude that most of the investment impact from the agreements comes from their non-trade provisions. 20

Medvedev (2006) notes that the Adams et al. (2003) study, is useful in determining the net FDI effect of a particular PTA, [but it] cannot establish the impact of preferential liberalization on a net FDI position of a particular country. Jaumotte (2004) tries to address this concern in a study of developing countries from 1980 to 1999. She finds that the market size due to the PTA at the beginning of a period has a significant and positive effect on the level of FDI stock at the end of the period. She concludes that, on the whole, her evidence is insufficient to support a claim that PTAs are desirable since it is unclear whether the costs associated with trade diversion are outweighed by the benefits of increased FDI. Lederman et al. (2005), in a study of aggregate FDI flows from 1980 to 2000, find that the expectation of joining a PTA in the next two years is associated with a one-third increase in FDI flows. The most recent studies on the association between PTAs and FDI flows have expanded on the original studies through changes in model specifications and by including more countries and PTAs. Hicks (2007), for instance, improves the model specification by including a variable to account for the type of PTA: preferential trade area, free trade area, customs union, common market, monetary union, single market, or economic and monetary union. He shows that, across a small sample of 25 PTAs, higher levels of PTA economic scope (i.e., the number of financial, fiscal, and monetary stipulations the PTA can enforce upon its member countries) and independence (i.e., the legitimate supranational power of the PTA) are associated with higher inward FDI flows. 21

Medvedev (2006) provides perhaps the most comprehensive study to date on the correlation between PTAs and increased FDI flows. He builds upon the Lederman et al. (2005) model, using a panel of 143 countries from 1980 to 2003 and examining, by far, the most comprehensive list of trade agreements of any study in the literature. His baseline model, which serves as a starting point for this paper, includes FDI flows as the dependent variable and an array of time-specific independent variables: GDP, GDP growth, a measure of the country s openness to trade, inflation, the change in the exchange rate, the level of world FDI, the world s GDP growth, the combined GDP of the countries involved in the PTA when in force, and the distance between the countries in the PTA. This model leads to several important conclusions: 1) FDI flows from PTAs increase with the size of PTA members and their proximity to the host country; 2) the relationship between PTAs and FDI flows was strongly driven by the developing countries in the model; and 3) the link between PTAs and FDI flows is strongest in the late 1990s and early 2000s, the period when most of the new deep integration agreements were signed. This paper contributes to the existing literature in several important ways. First, it not only studies PTAs, but also WTO accessions, and tries to determine whether becoming a member of the WTO is a greater sign of policy reform and stability than implementing a PTA, and, therefore, a better signal for investors to send FDI to that country. Second, it analyzes the partial effect of two particularly salient institutional variables in primary enrollment ratios and political constraints indices. FDI could be 22

attracted to countries with low human capital levels if the industry involved is laborintensive, or FDI could be attracted to countries with high human capital levels if the industry involved is capital-intensive. The hypothesis as to the partial effect of political constraints on FDI flows is similarly ambiguous because a politically constrained society could either signal strong protection of individual and property rights, or a society with low levels of political constraints could signal a better opportunity for investors to influence legislation in a manner that would be advantageous to their investments. Third, it attempts to determine whether different types of trade agreements preferential trade, free trade, customs unions, common markets, and other economic integration agreements with trade provisions are associated with different FDI outcomes. The initial conjecture as to the magnitude of this relationship would be that the higher the degree of integration, the more likely it is that countries will commit to liberal reforms to attract FDI. Finally, this paper examines the impact of FDI flows specifically to and originating from the member countries involved. Past work in this area has examined the impact of trade agreements on overall FDI flows, but it would also be useful to know whether most of the impact on FDI flows could be attributed to the member countries of the trade agreement, since investors in those countries would have the most intimate knowledge of the investment climate and the effect of the trade agreement on it. 23

Conceptual Model This paper seeks to determine whether, holding a number of country and globallyspecific variables equal, implementation of a preferential trading agreement (PTA) is associated with an increase in foreign direct investment (FDI) flows to the member countries of the PTA. The dependent variable in question is FDI flows, while the independent variable of most interest is the establishment of a trade agreement. PTAs Could Cause FDI There are many reasons to believe that signing a trade agreement will have a positive impact on FDI flows for countries entering that agreement. Investors may be reassured that the economic landscape in the host country is stable, and that their investment has a much lower chance to be institutionally discriminated against by the host country, because the host has established closer ties with investor countries, often including specific investment provisions in the agreement, in part under the assumption of closer economic integration (Büthe and Milner, 2004). U.S. bureaucracies that are directly impacted by trade agreements have voiced support for them by arguing they will have a positive impact on investment. Thus, the U.S. Chamber of Commerce has stated that [the U.S.-Chile Free Trade Agreement] will immediately boost bilateral trade and investment opportunities, to the ultimate benefit of both nations (USTR, 2002). Similarly, the Office of the United States Trade Representative states in their one-page document supporting the North American Free Trade Agreement (NAFTA) that NAFTA has benefited the United States because trade and investment flows have substantially 24

increased and U.S. non-residential fixed, or business, investment has risen by 107 percent [from 1993 to 2007], compared to a 45 percent increase between 1980 and 1993 (USTR, 2007). A growing literature supports the correlation between trade agreements and increased FDI flows. For example, Blömstrom and Kokko (1997), Adams, Gali, and McGuire (2003), te Velde and Fahnbuulleh (2003), Jaumotte (2004), Lederman, Maloney, and Serven (2005), Medvedev (2006), Hicks (2007), MacDermott (2007), and Park and Park (2007), using different methodologies and looking at different trade agreements, have all established a correlation between trade agreements and FDI flows. Of course, in order to understand the effect of PTAs on country FDI, it would be misleading and inappropriate to focus an analysis solely on the relationship between those two variables. There are three potential scenarios that could produce such a relationship aside from the relationship between PTAs and FDI. First, country-specific observable characteristics could have relationships with both PTAs and FDI. Second, unobserved country traits could cause both PTAs and FDI. Third, FDI itself could cause PTAs to form. Conditions that Could Cause both PTAs and FDI Openness of the Economy The overall openness of an economy to international trade could confound the simple relationship between PTAs and FDI, because each could be correlated with openness. The a priori expectations of the overall effect of the openness of an economy on FDI flows are somewhat 25

ambiguous. In terms of the two types of FDI, as trade openness in an economy increases, horizontal FDI could fall because most companies could use trade to enter markets instead of investment. However, as trade openness in an economy increases, vertical FDI could increase as companies find it profitable to produce components and final goods in other countries (MacDermott, 2007). Whether the effects on vertical FDI are stronger than the effects on horizontal FDI, or vice versa, could depend on whether the host country is developed or developing. Andriamananjara (2003) notes that, as a country decides whether to enter a PTA, it faces a trade-off between opening its own economy (increased competition) and gaining access to the PTA s market (preferential access). The market access gain is always larger as long as the aggregate size of the PTA s market is larger than the market of the prospective member; therefore, increasing openness should lead to more PTAs. Exchange Rate The exchange rate of a country can influence FDI decisions by affecting the prices of host country assets, the value of transferred profits, and the competitiveness of foreign affiliate exports (World Investment Report, 1998). Therefore, as a country s exchange rate depreciates FDI flows should increase to that country because the currency depreciation has effectively reduced production costs for foreign investors (Medvedev, 2006). Exchange rate stability has also been linked to the establishment of PTAs. Shin and Wang (2003) point out that exchange rate stability and its implications for a free trade agreement are 26

so commonly documented in the literature that the two are assumed to be subsidiary. However, the authors also indicate that certain trade agreements can lead to increased exchange rate stability and integration (e.g., the European Common Market), while others, typically bilateral agreements, can make exchange rates less stable. This inconsistency brings up an important endogeneity issue that could be present in several key variables or conditions in the core model for this paper. GATT/WTO Membership Unlike previously mentioned factors that impact FDI flows, whether or not a host country is a member of the GATT/WTO is only included in one of the previously mentioned studies. If a country belongs to the GATT/WTO, it could have access to many world markets as a source for inputs and outputs. Also, much like the potential effects of signing a PTA, being a member of GATT/WTO can provide a signal to investors that the country s government is open to world markets and liberalization policies (Büthe and Milner, 2004). Therefore, acceding to the GATT/WTO should be positively correlated with FDI flows. There is also much debate in trade literature as to whether or not joining the WTO makes a country more or less likely to pursue trade agreements outside of the WTO s purview. Algeria, for instance, joined the WTO at least in part in order to conclude a free trade agreement with the United States. The country was able to use its new trade contacts and institutional 27

knowledge as a WTO member to negotiate an agreement important to its economy on a more equal footing (Kennedy, 2004). Shocks Many additional factors, which are usually much less predictable, could simultaneously be causing both trade agreements and FDI flows. A sudden change to a political regime that is less friendly to FDI and/or trade, for example, could have a detrimental impact on both outcomes. Li and Resnick (2003), for example, state that political regime volatility will increase investor uncertainty about the host country s economic policies and, more specifically, FDI policies. Countries have also been hindered from joining trade agreements by inter-state tension or conflicts. In South Asia, for example, country leaders hoped to achieve economic integration through the South Asian Association for Regional Cooperation. But continued tensions both within individual countries and between countries have blocked that integration movement (Brown et al.). Indeed, this potential effect could have multiple directions, considering that certain trade agreements are signed specifically to design conflict out of intercountry relations, with the European Union serving as an obvious example. Conditions that Could Be Caused by FDI and PTAs Governance/Political Environment Some studies try to capture these empirically difficult indicators that could cause FDI, while others note its importance, but exclude them from their models due to practical concerns. 28

Institutions including tax systems, easiness to create a company, lack of corruption, transparency, contract law, security of property rights, and efficiency of justice and prudential standards can all be major determinants of FDI flows (Bénassy-Quéré, Coupet, and Mayer, 2005). As institutions worsen, costs of doing business increase, predictably diminishing FDI flows (Blonigen, 2005). However, these conditions have also been shown to be affected by signing trade agreements. For example, many recent trade agreements, especially those with a developed and developing country, include provisions that are meant to reduce corruption and increase transparency (e.g., the Canada-Peru FTA), or formally secure a substantial system of property rights (e.g., the Australia-Chile FTA). GDP and GDP Growth Rate Yet another set of characteristics, which could either be a cause of FDI and PTAs or be affected by FDI and PTAs, are a country s GDP and GDP growth rate. All of the previously mentioned studies in this field have at least one variable that takes into account some form of the host country s GDP level. Larger markets in terms of GDP are more likely to attract FDI because they have a greater expected stream of future returns. Similarly, GDP growth can induce FDI flows to a country because foreign investors are attracted to future market opportunities (Li and Resnick, 2003). Horizontal FDI, the most common type, will more likely flow to developed economies with high levels of GDP because firms desire markets with similar development levels to purchase their goods. Contrarily, vertical FDI will more likely flow to 29

developing economies with low levels of GDP where labor is cheap for the production process (Hicks, 2007). Thus, GDP should hypothetically be positively associated with FDI in developed countries and negatively associated with FDI in developing countries. FDI, however, can also be a cause of GDP growth. Carkovic and Levine (2004) summarize macroeconomic evidence on the impact of FDI on economic growth, while ultimately disagreeing with the results in some authors econometric studies. Borensztein et al. (1998), for instance, find that FDI positively effects growth in countries with highly educated workforces. Blomström et al. (1994) find that FDI has positive effects on growth when countries are sufficiently wealthy. Finally, Balasubramanyam et al. (1996) find a similar effect when a country has trade openness. The history of trade agreements has also clearly shown a positive impact on GDP growth. NAFTA, for example, has had a small but positive impact on American, Mexican, and Canadian GDP, averaging somewhere between $300 million and $2.1 billion per year for America in the first eight years of its existence (CBO, 2003). The murkiness of the relationship between FDI and GDP growth is best illustrated by Chowdhury and Mavrotas (2006), who find that GDP causes FDI in the case of Chile, while there is strong evidence of bi-directional causality between the two variables in Malaysia and Thailand. 30