1

INVESTMENT POLICY MONITOR

Note: This report can be freely cited provided appropriate acknowledgement is given to UNC-TAD and UNCTAD’s website is mentioned (www.unctad.org/diae).

IntroductionThis Monitor

A PERIODIC REPORt by thE UNCtAD sECREtARIAt

U n i t e d n at i o n s C o n f e r e n C e o n t r a d e a n d d e v e l o p m e n t

1

September 2014

The ImpacT of InTernaTIonal InvesTmenT agreemenTs on foreIgn DIrecT InvesTmenT: an overvIew of

empIrIcal sTuDIes 1998–2014

IIA ISSUES NOTE – Working Draft

Working Draft for peer review on the investmentpolicyhub platform. Comments welcome. Citation possible upon request.

This issues note has not been formally edited. Advance copy.

IIa Issues noTe

This working draft for a forthcoming IIA Issues Note invites investment and development stakeholders to engage in this topical debate. Posted on the investment policy hub’s discussion platform, it encourages academics, policy makers and the community at large to help improve our common understanding on the empirical evidence and underlying policy issues. Lending itself to a multi-disciplinary approach to research and policymaking on investment and development issues and calling for further research across a wide range of methodologies, this process will also feed into the Multi-disciplinary Academic Conference that will take place during the forthcoming World Investment Forum (WIF), 13-16 October in Geneva. “Investing in sustainable development”, the WIF’s overall guiding motto, is the background against which this online peer-review process is taking place.

Key messages

• Over the years, numerous empirical studies have assessed the impact of international investment agreements (IIAs), including bilateral investment treaties (BITs), on foreign direct investment (FDI) – with mixed results. A policy debate is now underway to reappraise previous findings and unsolved questions.

• An important consideration for the policy debate is the ultimate function of IIAs with respect to countries’ overall development strategies. Attracting FDI is neither the prime - nor the only - role of IIAs.

• The point of departure for present research and policy analysis is that IIAs are one of several determinants of FDI, and their importance is likely to be contingent on other variables. Other host-country policy areas are also significant, and so are business facilitation initiatives and economic determinants. Since IIAs play a complementary role among several determinants, they cannot substitute for sound domestic policies, regulatory and institutional frameworks.

• Existing empirical studies of IIAs’ impact on FDI provide heterogeneous results and have some limitations because of, among others, data and methodological challenges. The majority of studies conclude that IIAs have a positive impact on FDI. An empirical correlation does not necessarily imply causation, but most recent studies that address the so-called endogeneity problem are able to

2

establish a positive causal relationship between IIAs and FDI. More nuanced recent research finds that the content of IIAs matters: the FDI impact of IIAs is dependent on the presence of certain substantive treaty provisions. On the other hand, some empirical studies find no effect of IIAs on FDI flows.

• Across the board, empirical studies point to the importance of host country conditions. The quality of institutions, the level of political risk, or the development of the financial sector all influence companies’ investment decisions.

• Properly understanding the potential impact of IIAs on FDI is important for defining their role in countries’ investment policies and overall development strategies. Econometric studies can help, but also have limitations. Moreover, prominent counterfactuals (i.e. investment relationships that exist without being covered by IIAs) suggest that legal instruments’ influence on economic matters are limited and that other determinants, in particular the economic ones, are more important. Still, the question of whether an IIA would improve such an investment relationship remains open.

• Finally, the role of IIAs has to be put into the broader context of countries’ efforts to attract and benefit from FDI, with the ultimate objective to promote sustainable development. It is therefore important to consider the challenges that IIAs can give rise to, including with respect to potential constraints on policy space or exposure to investment litigation.

• What matters is the creation of a new generation of investment policies that places inclusive growth and sustainable development at the heart of efforts to attract and benefit from investment. UNCTAD’s Investment Policy Framework for Sustainable Development (IPFSD) can assist policy makers in doing so.

IntroductionThe regime of international investment agreements (IIAs) is undergoing a period of reflection, review and reform. In this context, an important policy debate is underway on the role of IIAs in promoting foreign direct investment (FDI). Thus, an important consideration is the ultimate function of IIAs with respect to countries’ overall development strategies. Attracting FDI is neither the prime – nor the only – role of IIAs. Instead, the key function of IIAs is to contribute to predictability, stability and transparency in investment relations. In this regard, IIAs can help improve countries’ regulatory and institutional frameworks, including by adding an international dimension to them. IIAs can reduce risks for foreign investors and, more generally, signal a better investment climate. Through all of this, IIAs can help facilitate cross-border investment flows, which, if managed properly, can help achieve sustainable development objectives.

The role of IIAs in promoting FDI has been the subject of a number of empirical studies. There are a significant number of econometric studies that address the question of the effectiveness of IIAs, and especially of bilateral investment treaties (BITs).

The purpose of this IIA Issues Note is twofold. First, to put the question of IIAs’ impact on FDI into the broader perspective of what determines FDI flows; second, to present an overview of the diverse empirical literature on the relationship between IIAs and FDI (primarily econometric, but also touching on other methodologies), spanning papers published from the late 1990s to the present.

In so doing, this IIA Issues Note aims to inform policy makers, academia, and other investment and development stakeholders about the current state of this on-going debate, noting that further research – across a wide range of methodologies – is warranted.

3

Host Country determinants Type of FDI classified Principal economic determinantsby motives of TNCs in host countries

I. Policy framework for FDI• Economic, political and social stability• Good governance• Policies on functioning and structure

of markets(especially competition, M&A and simple, transparent report-ing standards in line with international practise)

• Protection of property rights (including intellectual property)

• Industrial and regional policies; devel-opment of competitive clusters

• Trade policy (tariffs and non-tariff bar-riers) and stable exchange rates

• International Investment Agree-ments

II. Economic determinants

III. Business facilitation• Investment promotion• Investment incentives• Reduction of hassles costs • Availability of one-stop shop services• Provision of social amenities • Provision of after-investment services

A. Market-seeking • Per capita income• Market size • Market growth• Access to regional/global

markets

B. Natural resource-seeking• Access to raw materials

C. Efficiency-seeking• Different comparative advan-

tages of countries• Better deployment of global

resources

D. Strategic asset-seeking• Access to new competitive

advantages• Availability of and access to

skilled labour• Strategic infrastructure (e.g.

oil pipelines, power grids)

Table 1. Host country determinants of FDI

The determinants of fDI and the role of IIas

The determinants of FDIExisting research and policy analysis suggest that FDI flows are influenced by a wide range of factors. These factors include a country’s policy framework, its economic attractiveness (which depends on the motives of investors), and the business facilitation framework in place (Table 1).

The policy framework for FDI comprises all host-country policy areas that may be relevant to a foreign investor. These vary from one investment to the other. IIAs are only one element within the overall policy framework. While they are a part of the policy framework explicitly targeting FDI issues, other policy areas may be even more critical for a foreign company’s investment decision.

Economic determinants differ according to the main motive of an investor: market-seeking FDI is driven by considerations of market size and growth, access to regional or global markets, or specific market structure characteristics. Resource/asset-seeking FDI is aimed at gaining access to raw material, skilled and/or unskilled labour, technology and other created assets, and infrastructure. Efficiency-seeking FDI is undertaken to rationalize elements of a transnational corporation’s (TNC) global value chain (GVC) by accessing inputs (goods and services, including for trade and communication as well as skilled and unskilled labour) at lower costs and making use of economic cooperation and integration agreements for international production networks.

Finally, business facilitation measures (including promotional activities, incentive provisions, and other measures catering to TNCs’ needs) may tilt the balance in favour of a location once other preconditions are in place.

Source: UNCTAD, 1998a and 2010.

4

The IIA regime as a determinant of FDIThe expected role of IIAs as determinants of FDI includes (UNCTAD, 2012):

• adding an international dimension to investment protection and by fostering stability, predictability and transparency, reinforce investor confidence and thus promote both investment and trade flows.

• promoting investment in other ways beyond granting investor protection. Some IIAs include commitments on the part of home countries to promote outward investment or to engage in collaborative initiatives for this purpose (although this is currently a small minority of treaties).

• helping to build and advertise a more attractive investment climate. By establishing international commitments, they can foster good governance and facilitate or support domestic reforms.

However, IIAs alone  cannot  turn a weak domestic investment climate into a strong one and they cannot guarantee the inflow of foreign investment. There is no mono-causal link between the conclusion of an IIA and FDI inflows; IIAs play a complementary role among many determinants that drive firms’ investment decisions. Most importantly, IIAs cannot be a substitute for domestic policies and a sound national regulatory framework for investment.

Another important point relates to the type of IIAs a country has concluded. While BITs are self-standing and explicitly focused on issues of foreign investment, there are broader economic agreements where investment disciplines are just one element in a comprehensive treaty that encompasses trade in goods and services, and covers other aspects of economic cooperation (“other IIAs”). The latter can encourage FDI in additional ways, as they not only protect (and possibly provide access for) investment, but may also dismantle trade barriers, facilitate integration into GVCs, and lead to the creation of a larger market. These “other IIAs” with substantive investment provisions not only change the policy framework, but also affect the economic determinants of FDI. Currently, such treaties account for a very small percentage of all IIAs.

empirical studies on the relationship between IIas and fDIThe impact of IIAs on FDI flows has been the subject of empirical analyses for many years. Most of these studies focus on BITs. A comprehensive review of the literature since the late 1990s reveals two fundamental points (see the full analysis in the Annex).

• First, the majority of studies find a positive impact of IIAs on FDI. This also includes those studies that address the endogeneity problem (i.e. concerns about the direction of causality between IIAs and FDI, Box 1).

• Second, more nuanced recent research finds that the content of IIAs matters: IIAs positively influence FDI flows, provided that they include certain substantive provisions.

As ever more investment-related provisions are integrated into broader economic agreements, their coverage of a wide set of pertinent FDI determinants may be even more effective than BITs in attracting FDI; and research on them is growing accordingly.

The majority of econometric studies find a positive correlation between the presence of BITs and “other IIAs” and FDI. In their seminal paper, Neumayer and Spess (2005) used panel data for 119 countries over the period 1970 to 2001. They showed a positive effect of BITs on FDI inflows that is consistent and robust across various model specifications. The impact of BITs was found to be conditional on countries’ institutional quality. Other early studies produced

5

similar results: Büthe and Milner (2004) used a sample of 122 developing countries from 1970 to 2000 and found a “predicted positive, statistically and substantially significant correlation between BITs and subsequent inward FDI into developing countries”. Similarly, Egger and Pfaffermayr (2004) found that “BITs exert a positive and significant effect on outward FDI”. More recent studies also demonstrate a positive impact of BITs on FDI: Berger et al. (2013) in their sample of 28 home and 83 host countries found a significant positive impact, as did Tortian (2012) who looked at FDI inflows into 20 Southeast European and Central Asian countries. A positive impact of BITs on FDI is also reported by Oh and Fratianni (2010), Guerin (2010), Kerner (2009), Banga (2008), Siegmann (2008), Tobin and Rose-Ackerman (2006) and Grosse and Trevino (2005). Some studies nuance these results. For example, Salacuse and Sullivan (2005) suggest that United States (US) BITs are more likely to induce FDI inflows than those concluded by other OECD countries. Tobin and Rose-Ackerman (2011) concluded that “BITs do attract FDI to developing countries, but … cannot entirely substitute for an otherwise weak investment environment”. According to Pinto et al. (2010) ratified BITs have a significant and sizable effect on FDI flows but the impact is not stable and dissipates over time. Desbordes and Vicard (2009) mention that “the effect of BITs crucially depends on the quality of relations between the signatory countries”.

Although an empirical correlation does not necessarily imply causation, most recent studies that address endogeneity find a positive causal relationship between IIAs and FDI. The causal relationship between IIAs and FDI might theoretically run in both directions. Not only may IIAs attract FDI, but countries may also sign IIAs with other countries with which they already have a strong FDI relationship. Such reverse causality, as well as the potential for omitted variables, creates endogeneity problems, which more recent empirical literature has sought to address (Box 1). Aisbett (2009) notes that papers that ignore reverse causality find, in general, greater effects on FDI. However, recent papers confirm that the positive relationship between BITs and FDI holds even if endogeneity is taken into account. For instance, Busse, Königer, and Nunnenkamp (2010) employ a gravity-type methodology and various model specifications, including an instrumental variable approach, and find that BITs do promote FDI flows to developing countries. Berger et al. (2013), also using a gravity model, covering the 1978-2004 period and 28 home and 83 host countries, come to the same conclusions. Egger and Merlo (2007), using an unbalanced panel covering 24 home and 28 host countries between 1980 and 2001, find a strong contribution in the first years of an IIA’s existence. Colen et al. (2014), covering FDI stocks for 13 countries in the Commonwealth of Independent States (CIS) and in Central and Eastern Europe, clearly show a positive impact, but this impact varies by sector. Nevertheless, some papers dissent: e.g. the Peinhardt and Allee (2012) study on the impact of US BITs and Preferential Economic Agreements finds that very few countries have witnessed increased investment flows after signing such deals with the US, once endogeneity is accounted for. Similarly Aisbett (2009) tested OECD countries’ investment flows and found that the “initial strong correlation between BITs and investment flows is not robust controlling for selection into BIT participation” (i.e. controlling for endogeneity).

Some studies found little – or no – effects of BITs on FDI. UNCTAD’s (1998b) econometric study of a cross-sectional time-series model of the determinants of bilateral FDI inflows in 72 host countries over 23 years found that the relationship between BITs and FDI was weak – BITs could be expected to only “marginally increase” FDI. Hallward-Driemeier (2003) analyzed 20 years of bilateral FDI flows from OECD countries to developing countries with respect to several dependent variables (absolute amount of FDI, the ratio of FDI to host country GDP and the share of host country FDI in total FDI outflows of a home country) and found little evidence that BITs have stimulated additional FDI. Tobin and Rose-Ackerman

6

(2003) examined bilateral FDI flows between the US and 54 developing countries and found that, overall, the number of BITs has little impact on a country’s ability to attract FDI. This result is also supported by Peinhardt and Allee (2008) and Gallagher and Birch (2006).

Box 1. empirical challenges when analyzing the relationship between IIas and fDI

Correlation versus causation. The decision to enter into an IIA may be endogenous to FDI. Generally speaking, endogeneity makes it hard to properly distinguish causation (an additional IIA “causes” an increase of FDI) from simple correlation (an additional IIA “is empirically associated” with an increase in FDI). There are two sources of endogeneity: reverse causality and omitted variables. Reverse causality originates from the fact that increasing FDI flows may increase the probability that countries will sign an IIA: the motivation and the rationale for entering into an IIA is stronger if the two parties are already related (or expect to be related) by substantial investment relationships. The problem of omitted variables arises as a third exogenous variable acts simultaneously on FDI and IIAs, making their relationship spurious: for example, positive changes to a host country’s investment climate may at the same time stimulate more investment inflows and lead to a higher propensity to conclude BITs to ensure a safer legal framework for foreign investors.

Data limitations with respect to bilateral FDI. Most empirical studies rely on the classic gravity model, broadly adopted in the econometric literature on the determinants of FDI. In this class of models the dependent variable is represented by bilateral FDI (flows or stocks) between any two (source and host) countries. Covariates include a variable signaling the existence of a bilateral treaty between the countries of interest and a number of additional explanatory variables (acting as controls). These approaches face challenges from the limited availability of information on bilateral FDI. First, time-series of bilateral FDI data are available only for developed countries and a limited subset of developing countries. Second, focusing on developed countries leads to numerous “zeros” and “missing values” in the panel data (countries may have different reporting systems, for example reporting FDI only above a certain level, or may have started reporting bilateral FDI at different points in time).

These data limitations pose a number of challenges. First, data constraints limit the scope of the empirical analysis carried out. While many studies have investigated the impact of treaties between developed and developing countries (“North–South”), very limited attention has been given to the increasingly relevant “South–South” dimension for which bilateral investment data is poorer. Second, data limitations make it difficult to generalize results from a selected sample of reporting countries to a broader range of countries. Such a generalization may be affected by a selection bias due to systematic differences between countries that report and countries that do not report FDI data. Finally, data limitations give rise to technical challenges, e.g. those arising from the limited size of the statistical sample and the analytical treatment of zeros and missing values (especially in the presence of econometric models specified in logarithmic terms).

Also in terms of the explanatory variable, there are some major data challenges. Most studies employ an undifferentiated treatment of BITs, modeling them through a binary 0-1 variable. This approach ignores the fact that investment treaties contain different provisions, and their attractive capacity on FDI may depend on their content rather than their mere existence. The «black box» issue may be one of the key factors explaining the lack of conclusive evidence on the impact of BITs on FDI. Analytically the inclusion of the «content» of the BITs into the econometric analysis is not trivial as it requires a rigorous and comprehensive method to translate treatments’ provisions the legal framework into quantifiable and operational metrics.

These data and methodological challenges, as well as the very nature of econometric studies to work on the basis of a simplified description of a complex reality, make it difficult to draw policy conclusions from existing econometric studies on this matter.

Source: UNCTAD.

7

The early literature does not pay attention to the substance of BITs and other IIAs (i.e. their content) which can significantly affect their impact on FDI. Analytically, the inclusion of the “content” of IIAs into an econometric analysis is difficult, as it requires a way to translate the legal framework into quantifiable and operational metrics. Some studies have gone this extra mile. For example, the national treatment clause was found to be important in order for BITs to be effective (Berger et al., 2013). Berger et al. (2011) show that any positive FDI effects of BITs can be attributed to ISDS provisions; however the effectiveness of this relationship remains elusive due to its sensitiveness to the model specification, in particular to the inclusion of Central and Eastern European countries. In addition, and not sufficiently empirically tested yet, the overall nature of IIAs may matter, in particular whether a treaty is limited to post-establishment treatment of investments, or also includes guarantees for the establishment stage. Pre-establishment IIAs, especially if they involve liberalization of access conditions for investors (i.e. if they result in liberalization at the domestic level), are more likely to lead to an increase of FDI flows. Currently, only about 10 per cent of all IIAs provide for pre-establishment protection.

The status of an IIA (e.g. whether and when it moved from signature to ratification and entry into force) also influences their effect on FDI. To have an impact on FDI, an IIA need not merely be signed but must also enter into force, after being ratified by both countries. By the end of 2013, about 77 per cent of concluded IIAs had entered into force. Haftel (2010) shows that only ratified BITs have a statistically significant effect on FDI. The effect of BITs on FDI may also not be stable over time, with the impact being stronger in the interval immediately after an agreement’s entry into force, and dissipating over time. Egger and Merlo (2007) and Pinto et al. (2010) found that the effect of a BIT on FDI inflows is concentrated in the first years after its entry into force and is much weaker later.

The impact of “other IIAs”, especially preferential trade and investment agreements (PTIAs), on FDI flows is stronger than the impact of BITs. The impact of preferential trade and investment agreements (PTIAs) on FDI inflows is generally found to be positive (Medvedev, 2012; Büthe and Milner, 2014; World Bank, 2005; Banga, 2003; and Dee and Gali, 2003; see also UNCTAD, 2009 and Te Velde and Bezemer, 2004, for a synopsis of studies). That is even more so when agreements are ratified and include ISDS mechanisms (Büthe and Milner, 2014). Berger et al. (2013) found a positive impact arising from regional trade agreements (RTAs) in cases where liberal admission rules were included. Peinhardt and Allee (2012), on the other hand, found little influence from US BITs and preferential economic agreements on US FDI flows.

An IIA might also have a higher impact on FDI flows if it forms part of a country’s broader effort to attract FDI through other investment promotion activities (e.g. through investment promotion and facilitation measures etc.) or complementary regulatory and institutional reform (e.g. improving governance practices, reducing corruption, or building institutions).

However, it is difficult to disentangle the impact that various elements of PTIAs have on FDI. Lesher and Miroudot (2007) tried to overcome this through the construction of a composite index on the extensiveness of investment provisions in RTAs, and found this measure to have a significantly positive effect on FDI flows. Overall, the evidence, albeit based on a limited number of studies, suggests that the impact on FDI tends to be stronger for PTIAs than for BITs. This corresponds to the method by which the impact of IIAs on FDI is theorized: PTIAs influence a wider range of policy and economic determinants of FDI than BITs.

The host country environment and other factors shape the effectiveness of IIAs. The effectiveness of an IIA may also vary considerably depending on other host country factors, such as governance, institutional quality and political risk.

8

Neumayer and Spess (2005) and Siegmann (2008), in line with earlier results of Hallward-Driemeier (2003), show that the impact of BITs on FDI can depend on the institutional quality in the host country, while Yackee (2007) finds aspects of institutional quality – referring to them as part of political risk – to have the opposite effect. The latter may be in line with the results of Tobin and Rose-Ackerman (2003) which show that BITs are more effective in attracting FDI in relatively riskier countries. Similarly, Tortian (2012) shows that BITs do not play a strong role in FDI attraction where financial systems are well developed. Allee and Peinhardt (2011) find that the contribution of BITs to FDI attraction turns negative if countries are challenged through ISDS procedures and even more strongly so following a case in which the government loses.

Other approaches have been used to examine the impact of IIAs on FDI. Surveys, firm-level data, and transaction-level data produce mixed results. UNCTAD’s 2007 survey found that for an overwhelming majority (70 per cent of the surveyed TNCs), IIAs play a role in the decision to invest in a host country (UNCTAD, 2007). Despite the difficulties to express the relationship between FDI and BITs in a single number, Bellak (2014) finds a statistically significant positive effect of 2 per cent increase of FDI (confidence interval: 1.4; 2.6 per cent), based on meta-analysis of empirical studies on inward and outward FDI flows and FDI stocks (309 observations). Bellak (2014) concludes that the effect derived is too small to be of practical relevance, given the volatility of FDI flows and stocks in general. Several studies use firm-level evidence of the impact of IIAs on individual firm’s investment decisions. For example, Egger and Merlo (2012) use firm-level data on the activity of German transnational corporations (TNCs) to monitor their investments’ responsiveness to BITs. They find that BITs have a positive effect on TNCs’ foreign investment activity by both raising the number of TNCs that are active in a particular host country and increasing the number of plants, FDI stocks and fixed assets per firm. Following the idea that IIAs reduce political risk for foreign investors, Jandhyala and Weiner (2012) analyzed transaction-level data on the sale of petroleum reserves in 45 countries and found that TNCs pay higher prices for these assets in countries where they are protected by IIAs, pointing towards a risk adjustment in asset prices. Yackee (2011) also tried to “indirectly” measure the impact of BITs on FDI by means of the impact of investment treaties on the perceived political riskiness of host countries. The results of a regression analysis did not show significance. Yackee also highlighted that providers of political risk insurance (PRI) do not take BITs into account when determining their insurance conditions; neither “do in-house counsel in large US corporations view BITs as playing a major role in their companies’ foreign investment decisions”. Similarly, Poulsen (2010) also reports that investors rarely inquire about existing BITs before investing and that “treaties have very little impact on PRI providers’ coverage and pricing policies”.

Some cases appear to demonstrate a “counter-factual” in reality. Several large developing countries, such as Brazil, China, India or South Africa, are major recipients of FDI flows, including from countries with which they do not have an IIA relationship. While this points to the importance of FDI determinants other than IIAs (notably economic determinants such as market size, market growth, and resource endowments), it also underlines the limitations that legal instruments display in terms of influencing economic matters.

However, the picture is, again, more nuanced. Looking, for example, at US FDI stock in these countries – all countries that do not have IIAs with the US – the percentage share of US FDI stock is considerably lower than the average share of US FDI stock globally.1 While this does not mean that FDI flows from the US would increase once and if an IIA with these countries would be concluded and

1 While the US share in global inward FDI stock is around 25 per cent, for these four countries the US share only ranges between 5 and 15 per cent.

9

enter into force, it points to the likelihood that IIAs could play a role in determining investment location decisions.

summary and outlook Properly understanding the potential impact of IIAs on FDI can help define their role in countries’ investment policies and development strategies. Equally, it is important to consider the challenges that IIAs give rise to, including their impact on policy space and the exposure to ISDS cases.

In terms of the former, the present IIA Issues Note has shown three things:

• First, locational decisions of foreign investments are determined by an array of factors, including those related to a host country’s policy framework, the economic determinants and the motives of investors, and business facilitation measures. IIAs are an important part of the investment policy framework: they can foster predictability, stability and transparency; reduce risks for investors and signal a better investment climate. However, IIAs alone cannot turn a weak domestic investment climate into a strong one and they cannot guarantee the inflow of foreign investment. IIAs are only one element within the overall policy framework.

• Second, reviewing the literature on the impact of IIAs on FDI reveals that empirical studies provide heterogeneous results, among others because of significant data and methodological challenges. The majority of studies find a positive impact of IIAs on FDI, with some studies establishing a causal relationship between the two. More nuanced research finds that the content of IIAs matters: IIAs positively influence FDI flows, provided that they include certain substantive provisions and guarantees. The status of IIAs is also important: treaties that are in force have a greater impact than those that have only been signed. While it remains problematic to draw policy conclusions from econometric studies, these studies confirm the broad pathways through which IIAs are theorized to influence FDI flows: from the perspective of investors, BITs and other IIAs provide stability (e.g. the literature refers to risk reduction as a determining factor), protect investors (e.g. ISDS and national treatment clauses are found to be important as a co-determinants) and, more generally, contribute to a better investment climate.

• Third, the impact of IIAs is also conditional on, or mediated by, a number of other factors beyond the treaties themselves. The key factors identified in the literature are the sector and industry in which the FDI takes place, the country of origin of investors, the governance and institutions pertaining in host countries, and the life-cycle of an IIA (e.g. the size of impact is highest immediately after entry into force and then tails off). While these factors are undoubtedly important, whether and how they are treated in studies varies (e.g. in terms of how they are defined and specified or the metrics used).

There are a number of relevant aspects that are not addressed as extensively in the literature. Further, in some cases it is not possible to examine them with studies based on macro-panel data. These aspects include the specific motives of investors or the entry of investors through contractual forms other than FDI (such as contract manufacturing), which are increasingly important in international investment and trade (e.g. regarding GVCs). More research would help clarify these and other unresolved topics related to the impact of IIAs on FDI flows. Such studies would need to use variables grounded in commonly agreed conceptual frameworks, with transparent metrics facilitating the comparison of models and results. Such future research would also utilize a range of complementary methodologies, to fully tease out the nuances and details of IIAs’ impact on FDI, and the relative importance of other factors.

10

Such more contextualized evidence is crucial for policy makers aiming to make best-possible choices when formulating their countries’ strategic approach to international engagement on investment. Such more contextualized evidence can help better embed international investment policy making in countries’ development strategy and aid in designing IIA provisions that maximize sustainable development benefits while minimizing risks.

These findings are particularly important at this critical juncture in treaty making, where the IIA regime is experiencing a period of reflection, review and reform. The challenges that IIAs have given rise to, including the concerns related to their development dimension, to the balance between the rights and obligations of investors and States, and to the systemic complexity of the IIA regime in general (UNCTAD, 2012) have led to a situation where almost all countries are parties to one or several IIAs, but many are dissatisfied with the current treaty regime (UNCTAD, 2014). Countries’ efforts to address these challenges reveal four broad paths of action: (i) maintain the status quo, e.g. largely refraining from changes in the way countries enter into new IIA commitments; (ii) implement selective adjustments, e.g. modifying models for future treaties but leaving the treaty core and the body of existing treaties largely untouched; (iii) disengage from the IIA regime, e.g. unilaterally terminating existing treaties or denouncing multilateral arbitration conventions; and, finally, (iv) undertake a systematic reform to address the IIA regime’s challenges in a holistic manner (UNCTAD, 2014).

Deciding which of these paths to pursue, and whether or not to have IIAs (and if so, in which shape or form) is a matter of choice for governments. Such decisions come with a number of trade-offs that will involve – as with any international treaty – giving up some policy space in return for benefits from the treaty partners. Such a decision is usually taken in a specific bilateral (less frequently, in a plurilateral) context, where the IIA is one piece of a broader picture determined by a variety of factors. These factors include, among others, the level of economic development of participating States, relative trade and investment positions, geopolitical factors, and the general approach to bilateral or regional economic cooperation. They can also include the role of IIAs as determinants of trade in the context of GVCs, and their usefulness for domestic (institutional) reform efforts. A country may be prepared to sacrifice some of its policy space and certain aspects of its sovereignty, if there is a prospect of larger benefits to be gained. Hence, in principle, countries would assess the costs and benefits of entering into an IIA in each case, and depending on the potential partner(s) involved.

In all of this, it is essential that governments take utmost care to ensure that these treaties contribute to, rather than impede, their countries’ overall development strategies, including with respect to industrial policy, social and environmental objectives. Success in attracting and benefiting from investment depends not only on investment policy “stricto sensu” (i.e. entry and establishment rules, treatment and protection), both at the national and international levels, but on a host of investment-related policy areas ranging from tax and industrial policies to trade to environmental and labour market policies. These policy areas interact with each other and there is consequently a need for a coherent overall approach to make them conducive to sustainable development (UNCTAD, 2012). UNCTAD’s Investment Policy Framework for Sustainable Development (IPFSD) can offer important guidance in this regard.

11

referencesAisbett, E. (2009). “Bilateral Investment Treaties and Foreign Direct Investment: Correlation

Versus Causation,” in: Sauvant, K. and L.E. Sachs (eds.), The Effects of Treaties on Foreign Direct Investment: Bilateral Investment Treaties, Double Taxation Treaties and Investment Flows. New York: Oxford University Press.

Allee, T. and C. Peinhardt (2011). “Contingent Credibility: The Impact of Investment Treaty Violations on Foreign Direct Investment”, International Organization, Vol. 65, No. 3, pp. 401–432.

Banga, R. (2003). “Impact of Government Policies and Investment Agreements on FDI Inflows”. Working Paper No. 116 (November). New Delhi: Indian Council for Research on International Economic Relations.

Banga, R. (2008). “Government Policies and FDI Inflows of Asian Developing Countries: Empirical Evidence”, in: Fanelli, J.M. and L. Squire (eds.), Economic Reform in Developing Countries Reach, Range, Reason. Cheltenham and Northampton: Edward Elgar, pp. 117-146.

Bellak, C. (2014). “Economic Impact of Investment Agreements”, Asian Development Bank, forthcoming.

Berger, A., M. Busse, P. Nunnenkamp and M. Roy (2011). “More Stringent BITs, Less Ambiguous Effects on FDI? Not a BIT!”, Economics Letter, Vol. 112, No. 3, pp. 270-272.

Berger, A., M. Busse, P. Nunnenkamp and M. Roy (2013). “Do Trade and Investment Agreements Lead to More FDI? Accounting for Key Provisions Inside the Black Box”, International Economics and Economic Policy, Vol. 10, No. 2, pp. 247-275.

Busse, M., J. Königer and P. Nunnenkamp (2010). “FDI Promotion through Bilateral Investment Treaties: More Than a Bit?”, Review of World Economics, Vol. 146, No. 1, pp. 147–177.

Büthe, T. and H.V. Milner (2004). “Bilateral Investment Treaties and Foreign Direct Investment: A Political Analysis.” A revised version of the paper presented at the Annual Meeting of the American Political Science Association (September, 2004), in: Sauvant, K.P. and L.E. Sachs (eds.) (2009), The Effects of Treaties on Foreign Direct Investment: Bilateral Investment Treaties, Double Taxation Treaties and Investment Flows. New York: Oxford University Press.

Büthe, T. and H.V. Milner (2014). “Foreign Direct Investment and Institutional Diversity in Trade Agreements: Credibility, Commitment, and Economic Flows in the Developing World, 1971-2007”, World Politics, Vol. 66, No. 1, pp. 88-122.

Colen, L., D. Persyn and A. Guariso (2014). “What Type of FDI Is Attracted By Bilateral Treaties?”, LICOS Discussion Paper Series No. 346/2014.

Dee, P. and J. Gali (2003). “The Trade and Investment Effects of Preferential Trading Agreements”, NBER Working Papers Series, Working Paper No. 10160. Cambridge, Massachusetts: National Bureau of Economic Research.

Desbordes, R. and V. Vicard (2009). “Foreign Direct Investment and Bilateral Investment Treaties: An International Political Perspective”, Journal of Comparative Economics, Vol. 37, No. 3, pp. 372–386.

Egger, P. and M. Pfaffermayr (2004). “The Impact of Bilateral Investment Treaties on Foreign Direct Investment”, Journal of Comparative Economics, Vol. 32, No. 4, pp. 788–804.

Egger, P. and V. Merlo (2007). “The Impact of Bilateral Investment Treaties on FDI Dynamics”, The World Economy, Vol. 30, No. 10, pp. 1536–1549.

Egger, P. and V. Merlo (2012). “BITs Bite: An Anatomy of the Impact of Bilateral Investment Treaties on Multinational Firms”, Scandinavian Journal of Economics, Vol. 114, No. 4, pp. 1240–1266.

12

Gallagher, K.P. and M.B.I. Birch (2006). “Do Investment Agreements Attract Investment? Evidence from Latin America”, The Journal of World Investment and Trade, Vol. 7, No. 6, pp. 961–974.

Grosse, R. and L.J. Trevino (2005). “New Institutional Economics and FDI Location in Central and Eastern Europe”, Management International Review, Vol. 45, No. 2, pp. 123–145.

Guerin, S. (2010). “Do the European Union’s Bilateral Investment Treaties Matter? The Way Forward After Lisbon”, CEPS Working Document No. 333.

Haftel, Y.Z. (2010). “Ratification Counts: US Investment Treaties and FDI Flows into Developing Countries”, Review of International Political Economy, Vol. 17, No. 2, pp. 348–377.

Hallward-Driemeier, M. (2003). “Do Bilateral Investment Treaties Attract FDI? Only a Bit and They Could Bite”, World Bank Policy Research Paper, WPS 3121. Washington, D.C. World Bank.

Jandhyala, S. and R. Weiner (2012). “Do International Investment Agreements Protect Investment?”, Petroleum Evidence, mimeo.

Kerner, A. (2009). “Why Should I Believe You? The Costs and Consequences of Bilateral Investment Treaties”, International Studies Quarterly, Vol. 53, No. 1, pp. 73-102.

Lesher, M. and S. Miroudot (2007). “The Economic Impact of Investment Provisions in Regional Trade Agreements”, Aussenwirtschaft, Vol. 62, No. 2, pp. 193-232.

Medvedev, D. (2012). “Beyond Trade: The Impact of Preferential Trade Agreements on FDI Inflows”, World Development, Vol. 40, No. 1, pp. 49–61.

Neumayer, E. and L. Spess (2005). “Do Bilateral Investment Treaties Increase Foreign Direct Investment to Developing Countries?”, World Development, Vol. 33, No. 10, pp. 1567–1585.

Oh, C.H. and M. Fratianni (2010). “Do Additional Bilateral Investment Treaties Boost Foreign Direct Investments?”, Indiana University, Kelley School of Business, Department of Business Economics and Public Policy, Working Papers.

Peinhardt, C. and T. Allee (2008). “The Costs of Treaty Participation and Their Effects on U.S. Foreign Direct Investment”, American Society for International Law’s International Economic Law Interest Group Meeting in Washington, DC, pp. 1-39.

Peinhardt, C. and T. Allee (2012). “Failure to Deliver: The Investment Effects of US Preferential Economic Agreements”, The World Economy, Vol. 35, No. 6, pp. 757-783.

Pinto, M.P., S.M. Pinto and N. Stier-Moses (2010). “Regulating Foreign Investment: A Study of the Properties of Bilateral Investment Regimes”, Paper prepared for presentation at the Annual Meeting of the International Political Economy Society, Cambridge, MA, November 12-13, mimeo, pp. 1-24.

Poulsen, L. (2010). “The Importance of BITs for Foreign Direct Investment and Political Risk Insurance: Revisiting the Evidence”, in: Sauvant, K. (eds.), Yearbook on International Investment Law & Policy 2009-2010. New York: Oxford University Press.

Salacuse, J.W. and N.P. Sullivan (2005). “Do BITS Really Work? An Evaluation of Bilateral Investment Treaties and Their Grand Bargain”, Harvard International Law Journal, Vol. 46, No. 1, pp. 67–130.

13

Siegmann, T. (2008). “The Impact of Bilateral Investment Treaties and Double Taxation Treaties on Foreign Direct Investments”, University of St. Gallen Law School Law and Economics Research Paper Series, Working Paper No. 2008-22.

Te Velde, D.E. and D. Bezemer (2003). “Regional Integration and Foreign Direct Investment in Developing Countries”, Overseas Development Institute, mimeo.

Tobin, J. and S. Rose-Ackerman (2003). “Foreign Direct Investment and the Business Environment in Developing Countries: The Impact of Bilateral Investment Treaties”, William Davidson Institute, Working Paper No. 587. Michigan: The University of Michigan Business School.

Tobin, J. and S. Rose-Ackerman (2006). “Bilateral Investment Treaties: Do They Stimulate Foreign Direct Investment?”, Yale University (June), mimeo.

Tobin, J. and S. Rose-Ackerman (2011). “When BITs Have Some Bite: The Political-Economic Environment for Bilateral Investment Treaties?”, The Review of International Organizations, Vol. 6, No. 1, pp. 1-32.

Tortian, A. (2012). “The Impact of Bilateral Investment Treaties and Financial Development on Foreign Direct Investment: Evidence from Eurasia”, Paper Submission for Armenian Economic Association Conference, 13-14, October, Yerevan, Armenia, Haigazian University, Beirut, Lebanon.

United Nations Conference on Trade and Development (UNCTAD) (1998a). World Investment Report 1998: Trends and Determinants. New York and Geneva: United Nations. United Nations publication, Sales No. E.98.II.D.5.

UNCTAD (1998b). Bilateral Investment Treaties in the Mid-1990s. New York and Geneva: United Nations. United Nations. United Nations publication, Sales No. E.98.II.D.8.

UNCTAD (2007). “Worldwide Survey of Foreign Affiliates”. Occasional Note. Geneva: United Nations.

UNCTAD (2009). The Role of International Investment Agreements in Attracting Foreign Direct Investment to Developing Countries. UNCTAD Series on International Investment Policies for Development. New York and Geneva: United Nations.

UNCTAD (2010). World Investment Report 2010: Investing in a Low-carbon Economy, New York and Geneva: United Nations. United Nations publication, Sales No. E.10.II.D.2.

UNCTAD (2012). World Investment Report 2012: Towards a New Generation of Investment Policies. New York and Geneva: United Nations. United Nations publication, Sales No. E.12.II.D.3.

UNCTAD (2014). World Investment Report 2014: Investing in the SDGs: An Action Plan. New York and Geneva: United Nations. United Nations publication, Sales No. E.14.II.D.1.

World Bank (2005). Global Economic Prospects. Trade, Regionalism and Development. Washington D.C.: World Bank.

Yackee, J.W. (2007). “Do Bits Really Work? Revisiting the Empirical Link Between Investment Treaties and Foreign Direct Investment”, Legal Studies Research Paper Series, Paper No. 1054. Wisconsin: University of Wisconsin Law School.

Yackee, J.W. (2011). “Do Bilateral Investment Treaties Promote Foreign Direct Investment? Some Hints from Alternative Evidence”, Virginia Journal of International Law, Vol. 51, No. 2, pp. 397-440.

14

A

NN

EX

: A s

umm

ary

of

sele

cted

eco

nom

etri

c st

udie

s o

n th

e im

pac

t o

f B

ITs

and

PT

IAs

on

FDI (

in c

hro

nolo

gic

al o

rder

)

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Büth

e an

d M

ilner

(201

4)FD

I/GDP

19

71-2

007

122

coun

tries

; 3,

067

coun

try-y

ear

obse

rvat

ions

•PT

IAs

in fo

rce

•ch

ange

in P

TIAs

in fo

rce

•di

sput

e se

ttlem

ent

wei

ghte

d PT

IAs

•W

orld

Tra

de O

rgan

izatio

n (W

TO)

mem

bers

hip

•tra

de/G

DP•

polit

ical

sta

bilit

y in

dica

tors

•m

arke

t size

•GD

P gr

owth

•ec

onom

ic d

evel

opm

ent

Erro

r cor

rect

ion

mod

el•

The

estim

ated

effe

ct o

f PTI

As in

forc

e is

hig

her t

han

for s

igne

d PT

IAs.

•Th

e es

timat

ed c

oeffi

cien

t of i

nves

t-m

ent-

wei

ghte

d PT

IAs

is p

ositi

ve a

nd

sign

ifica

nt.

•FD

I inc

reas

es o

ccur

for r

atifi

ed P

TIAs

, not

mer

ely

sign

ed P

TIAs

.

•PT

IAs

with

inve

stm

ent c

laus

es o

f dis

pute

-set

tlem

ent

mec

hani

sms

attra

ct m

ore

FDI.

Cole

n, P

ersy

n an

d Gu

aris

o (2

014)

Tota

l FDI

sto

cks

in a

cou

ntry

in a

sp

ecifi

c se

ctor

1994

-200

913

cou

ntrie

s in

the

form

er S

ovie

t Uni

on

and

Cent

ral a

nd

East

ern

Euro

pe

•Nu

mbe

r of B

ITs

by

coun

try•

Mon

thly

wag

es in

man

ufac

turin

g•

polit

ical

inst

itutio

nal q

ualit

y•

infla

tion

•tra

de o

penn

ess

•av

erag

e ca

pita

l sto

ck p

er fi

rm a

nd

capi

tal/l

abou

r rat

io

Fixe

d ef

fect

s m

odel

The

stud

y fin

ds th

at e

spec

ially

for i

n-ve

stm

ents

in th

e se

ctor

s of

util

ities

and

re

al e

stat

e, a

nd to

a le

sser

ext

ent f

or

bank

ing

and

min

ing,

BIT

s ha

ve a

robu

st

and

econ

omic

ally

sign

ifica

nt e

ffect

on

FDI s

tock

s. F

or fo

reig

n in

vest

men

ts in

m

anuf

actu

ring

and

serv

ices

, BIT

s se

em

not t

o pl

ay a

maj

or ro

le in

inve

stm

ent

deci

sion

s.

BITs

can

be

expe

cted

to b

e m

ost e

ffect

ive in

thos

e se

ctor

s w

ith la

rge

sunk

cos

ts, r

elat

ively

low

leve

ls o

f fir

m-s

peci

fic k

now

-how

and

sec

tors

that

are

pol

itica

lly

sens

itive

to fo

reig

n ow

ners

hip.

Berg

er, B

usse

, Nu

nnen

kam

p an

d Ro

y (2

013)

Bila

tera

l FDI

flow

s19

78-2

004

Hom

e: 2

8De

velo

ped

Host

: 83

deve

lopi

ng c

ount

ries

•Au

thor

-cre

ated

BIT

in

dex

•Re

al G

DP g

row

th

•in

flatio

n •

trade

ope

nnes

s •

DTT

•cu

rrenc

y

Grav

ity m

odel

for

dete

rmin

ants

of F

DIRT

As a

nd B

ITs

have

pos

itive

, sig

nific

ant

impa

ct o

n FD

I flow

s on

ly w

hen

NT p

rovi-

sion

s ar

e in

clud

ed.

The

pres

ence

of n

atio

nal t

reat

men

t pro

visio

ns h

as a

st

rong

, pos

itive

rela

tions

hip

with

FDI

, whe

reas

ISDS

m

echa

nism

s ap

pear

to p

lay

a m

uch

wea

ker r

ole

in d

eter

-m

inin

g FD

I.

Egge

r and

M

erlo

(201

2)Fi

rm le

vel d

ata

on

the

inte

rnat

iona

l ac

tivity

of G

erm

an

TNCs

in th

e ho

st

coun

tries

; tar

get

met

rics

of fo

reig

n ac

tivity

: num

ber o

f af

filia

tes,

em

-pl

oyee

s, a

sset

s,

turn

over

1996

-200

5Ge

rman

TNC

s; 8

6 ho

st c

ount

ries

•Ra

tifica

tion

of a

BIT

•si

gnat

ure

of a

BIT

•GD

P of

the

host

cou

ntry

•sk

illed

labo

ur e

ndow

men

ts•

capi

tal-l

abou

r rat

io•

stat

utor

y ta

x ra

te o

f the

hos

t cou

ntry

•pr

esen

ce o

f dou

ble

taxa

tion

trea

ties

•pr

esen

ce o

f a P

TIA

Pois

son

Fixe

d Ef

fect

s Qu

asi M

axim

um L

ikel

i-ho

od e

stim

atio

n

BITs

(bot

h si

gned

and

ratifi

ed) r

aise

the

num

ber o

f mul

tinat

iona

l firm

s th

at a

re

activ

e in

a p

artic

ular

hos

t cou

ntry

.

BITs

hav

e a

posi

tive

effe

ct o

n th

e nu

m-

ber o

f pla

nts

per fi

rm, a

s w

ell a

s on

FDI

st

ocks

and

fixe

d as

sets

per

firm

.

The

pres

ence

of B

ITs

incr

ease

s th

e in

tern

atio

nal a

ctivi

ty

of T

NCs

in th

e ho

st c

ount

ry.

Med

vede

v (2

012)

Net F

DI in

flow

s19

80-2

003

143

econ

omie

s•

Dum

my

for P

TIAs

join

ed•

GDP

•tra

de o

penn

ess

•GN

I•

GDP

grow

th•

infla

tion

•w

orld

FDI

inflo

ws

Pane

l est

imat

ion

PTIA

-rel

ated

var

iabl

es e

stab

lish

a po

sitiv

e lin

k be

twee

n ne

t FDI

flow

s an

d pr

efer

entia

l tra

de li

bera

lizat

ion.

PTIA

mem

bers

hip

is a

ssoc

iate

d w

ith a

pos

itive

cha

nge

in

net F

DI in

flow

s, a

nd th

e FD

I gai

ns a

re in

crea

sing

with

the

mar

ket s

ize o

f the

PTI

A pa

rtner

s an

d th

eir p

roxim

ity to

the

host

cou

ntry

.

Pein

hard

t and

Al

lee

(201

2)US

out

war

d FD

I flo

ws

1977

-200

7Un

ited

Stat

es a

s in

vest

orHo

st: 1

78 e

cono

-m

ies

•Du

mm

y fo

r yea

r PTI

A si

gned

•La

gged

FDI

Tim

e se

ries

inte

rven

-tio

n an

alys

isAg

reem

ents

rare

ly ha

ve a

sta

tistic

ally

sign

ifica

nt e

ffect

on

US F

DI fl

ows.

US B

ITs

and

PTIA

s ra

rely

resu

lt in

an

incr

ease

in F

DI.

15

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Torti

an (2

012)

Bila

tera

l out

war

d FD

I sto

ck19

92-2

010

Hom

e: 2

0 OE

CD

coun

tries

Host

: 20

Sout

h Ea

st E

urop

ean,

So

uth

Cauc

asus

an

d Ce

ntra

l Asi

an

coun

tries

•BI

T ex

iste

nce

•Re

gion

al e

cono

mic

agr

eem

ent d

um-

mie

s •

GDP

•gr

owth

•

infla

tion

•fin

anci

al d

evel

opm

ent

•cr

editw

orth

ines

s •

natu

ral r

esou

rce

endo

wm

ents

Pane

l est

imat

ion

The

BIT

varia

ble

is h

ighl

y si

gnifi

cant

in

all s

peci

ficat

ions

.

Regi

onal

eco

nom

ic a

gree

men

t var

iabl

e is

not

sig

nific

ant.

With

the

finan

cial

dep

th in

tera

ctio

n te

rm,

the

BIT

varia

ble

lose

s its

sig

nific

ance

.

Ratifi

catio

n of

BIT

s be

twee

n OE

CD a

nd E

URAS

IA c

ount

ries

exer

t a h

ighl

y si

gnifi

cant

pos

itive

effe

ct o

n bi

late

ral i

nwar

d FD

I.

The

Blac

k Se

a Ec

onom

ic C

o-op

erat

ion

RTA

had

no im

pact

on

intra

-reg

iona

l FDI

.

BITs

do

not p

lay

a pa

rticu

larly

stro

ng ro

le in

incr

easi

ng

FDI i

n th

e ca

se o

f wel

l-dev

elop

ed fi

nanc

ial s

yste

ms.

Alle

e an

d Pe

in-

hard

t (20

11)

Net F

DI in

flow

s19

84-2

007

Num

ber o

f BIT

s co

nclu

ded

•IC

SID

disp

utes

•

dem

ocra

cy•

prop

erty

righ

ts•

popu

latio

n •

GDP

per c

apita

•

GDP

•fin

anci

al o

penn

ess

•ex

chan

ge ra

te v

olat

ility

Pane

l dat

a, fi

xed

effe

cts

Gove

rnm

ents

acc

used

of i

nves

tmen

t tre

aty

viola

tions

bef

ore

ICSI

D ex

perie

nce

stat

istic

ally

and

subs

tant

ive re

duct

ions

in

FDI

.

All c

oeffi

cien

t est

imat

es fo

r “lo

st” I

CSID

di

sput

es a

re n

egat

ive a

nd s

igni

fican

t.

BITs

incr

ease

FDI

into

cou

ntrie

s th

at s

ign

them

, but

onl

y if

thos

e co

untri

es a

re n

ot s

ubse

quen

tly c

halle

nged

bef

ore

ICSI

D.

On th

e ot

her h

and,

gov

ernm

ents

suf

fer n

otab

le lo

sses

of

FDI

whe

n th

ey a

re ta

ken

befo

re IC

SID

and

suffe

r eve

n gr

eate

r los

ses

whe

n th

ey lo

se a

n IC

SID

disp

ute.

Berg

er, B

usse

, Nu

nnen

kam

p an

d Ro

y (2

011)

Bila

tera

l FDI

flow

s (e

xpre

ssed

as

shar

es);

thre

e ye

ars

aver

age

to

smoo

th a

nnua

l flu

ctua

tions

1978

-200

4Ho

me:

14

coun

tries

Host

: 83

deve

lopi

ng

econ

omie

s

•Ra

tified

BIT

with

out

effe

ctive

ISDS

•ra

tified

BIT

with

ef-

fect

ive IS

DS (t

he B

IT

cont

ains

stri

ct IS

DS p

ro-

visio

ns fo

r the

inve

stor

pr

otec

tion)

•Ho

st c

ount

ry G

DP a

nd G

DP g

row

th•

host

cou

ntry

infla

tion

•ho

st c

ount

ry o

penn

ess

to tr

ade

•di

ffere

nce

GDP

per c

apita

bet

wee

n ho

me

and

host

cou

ntry

•pr

esen

ce o

f DTT

•pr

esen

ce o

f com

mon

cur

renc

y•

pres

ence

of a

PTI

A

Gene

raliz

ed M

etho

ds

of M

omen

ts e

stim

a-to

r with

inst

rum

enta

l va

riabl

es

Any

posi

tive

FDI e

ffect

s of

BIT

s ca

n be

at

tribu

ted

to IS

DS p

rovis

ions

.

How

ever

the

effe

ctive

ness

of t

his

rela

tions

hip

rem

ains

elu

sive

due

to it

s se

nsiti

vene

ss to

the

mod

el s

peci

ficat

ion,

in

par

ticul

ar to

the

incl

usio

n of

Cen

tral

and

East

ern

Euro

pean

cou

ntrie

s.

The

resu

lts o

n th

e im

pact

of B

ITs

on F

DI a

re q

uite

elu

sive

. Th

e po

sitiv

e im

pact

of B

IT is

lim

ited

to B

ITs

with

effe

c-tiv

e IS

DS p

rovis

ions

. How

ever

the

effe

ctive

ness

of t

his

rela

tions

hip

rem

ains

elu

sive

due

to it

s se

nsiti

vene

ss to

th

e m

odel

spe

cific

atio

n, in

par

ticul

ar to

the

incl

usio

n of

Ce

ntra

l and

Eas

tern

Eur

opea

n co

untri

es.

Tobi

n an

d Ro

se-A

cker

-m

an (2

011)

Bila

tera

l FDI

dat

a19

84-2

007

97 h

ost c

ount

ries

•Ra

tified

BIT

sBI

Ts d

o at

tract

FDI

, pro

vided

the

host

cou

ntrie

s ha

ve

stro

ng d

omes

tic in

stitu

tions

.

Yack

ee (2

011)

Polit

ical

risk

ra

tings

of t

he h

ost

coun

tries

; thr

ee

diffe

rent

met

rics

of p

oliti

cal r

isk

Diffe

rent

tim

e pe

riods

ac

cord

ing

to

the

met

rics:

- 19

85-

2003

- 19

81

-200

3-

1983

- 19

97

Diffe

rent

set

s of

(h

ost)

deve

lopi

ng

econ

omie

s ac

cord

-in

g to

the

met

rics:

- 11

0-

35-

97

•Nu

mbe

r of «

stro

ng»

BITs

sig

ned

by th

e co

untry

of i

nter

est (

a BI

T is

cla

ssifi

ed a

s «s

trong

» if

it co

ntai

ns

the

stat

e’s

pre-

cons

ent

to in

vest

or-in

itiat

ed

arbi

tratio

n fo

r a w

ide

rang

e of

dis

pute

s)

•GD

P pe

r cap

ita•

infla

tion

rate

•du

mm

y va

riabl

e fo

r “m

ass

expr

opria

-to

r”•

stat

e’s

birth

yea

r•

dem

ocra

cy in

dica

tors

Two

mod

els

test

ed:

1. G

ener

alize

d Le

ast

Squa

res

with

Fixe

d Ef

fect

s2.

Pan

el C

orre

cted

St

anda

rd E

rror

Ther

e is

a s

tatis

tical

ly si

gnifi

cant

impa

ct

of B

ITs

on c

ount

ries’

risk

ratin

g on

ly fo

r on

e of

the

thre

e ris

k in

dica

tors

test

ed.

BITs

hav

e a

limite

d ca

paci

ty to

mod

ify th

e ris

k pr

ofile

of

host

cou

ntrie

s (a

nd th

us to

attr

act m

ore

inve

stm

ents

).

16

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Buss

e, K

önig

er

and

Nunn

en-

kam

p (2

010)

Bila

tera

l FDI

flow

s (e

xpre

ssed

as

shar

es);

thre

e ye

ars

aver

age

to

smoo

th a

nnua

l flu

ctua

tions

; ver

-si

on w

ith «

zero

» an

d ve

rsio

n w

ithou

t «ze

ro»

1978

-200

4Ho

me:

28

coun

tries

Host

: 83

deve

lopi

ng

econ

omie

s ex

clud

-in

g OF

Cs

•Pr

esen

ce o

f a ra

tified

BI

T be

twee

n th

e tw

o co

untri

es o

f int

eres

t

•Ho

st c

ount

ry G

DP a

nd G

DP g

row

th•

host

cou

ntry

infla

tion

•ho

st c

ount

ry o

penn

ess

•di

ffere

nce

GDP

per c

apita

bet

wee

n ho

me

and

host

cou

ntry

•pr

esen

ce o

f DTT

•pr

esen

ce o

f com

mon

cur

renc

y•

pres

ence

of a

PTI

A•

mea

sure

of o

vera

ll ca

pita

l ope

nnes

s

Diffe

rent

mod

el s

peci

fi-ca

tions

test

ed:

1. O

rdin

ary

Leas

t Sq

uare

s w

ith fi

xed

effe

cts

2. F

ixed

Effe

cts

Pois

-so

n Ps

eudo

Max

imum

Li

kelih

ood

3.

Gen

eral

ized

Met

hods

of

Mom

ents

est

ima-

tor w

ith in

stru

men

tal

varia

bles

Unde

r all

the

mod

el s

peci

ficat

ions

and

fo

r bot

h FD

I spe

cific

atio

ns (w

ith z

eros

an

d w

ithou

t zer

os),

the

coef

ficie

nt o

f the

BI

T va

riabl

e re

mai

ns p

ositi

ve a

nd s

igni

fi-ca

nt; i

n m

ost c

ases

at 5

or 1

% le

vel.

BITs

do

prom

ote

FDI fl

ows

to d

evel

opin

g co

untri

es a

nd

they

may

eve

n su

bstit

ute

for w

eak

dom

estic

inst

itutio

ns,

thou

gh p

roba

bly

not f

or u

nila

tera

l cap

ital a

ccou

nt li

ber-

aliza

tion.

Guer

in (2

010)

Bila

tera

l FDI

flow

s 19

92-2

004

Hom

e: 1

4 OE

CD

econ

omie

sHo

st: 2

5 ho

st

mid

dle

inco

me

emer

ging

mar

kets

•Pr

esen

ce o

f a B

IT b

e-tw

een

the

two

coun

tries

of

inte

rest

•Ho

me

and

host

cou

ntry

GDP

•so

urce

and

hos

t cou

ntry

GDP

per

ca

pita

Fixe

d-ef

fect

s es

tima-

tion

(with

cou

ntry

-pai

r fix

ed e

ffect

s an

d co

untry

-and

-tim

e du

mm

ies)

BITs

hav

e a

stat

istic

ally

sign

ifica

nt a

nd

posi

tive

impa

ct o

n FD

I flow

s be

twee

n th

e (1

4) O

ECD

coun

tries

and

thei

r BIT

pa

rtner

s.

The

estim

ated

coe

ffici

ent i

ndic

ates

th

at a

n ad

ditio

nal B

IT in

crea

ses

sour

ce

coun

try F

DI b

y 1.

3%.

BITs

do

prom

ote

FDI o

utflo

ws

from

OEC

D co

untri

es to

BIT

pa

rtner

s.

Hafte

l (20

10)

Bila

tera

l flow

s fro

m th

e Un

ited

Stat

es to

dev

elop

-in

g co

untri

es;

expr

esse

d as

pe

rcen

tage

on

the

GDP

of th

e ho

st

coun

try

1977

-200

4Th

e Un

ited

Stat

es

as in

vest

orHo

st: 1

20 d

evel

op-

ing

coun

tries

•Si

gned

BIT

s •

BITs

in fo

rce

•Ho

st c

ount

ry G

DP, G

DP g

row

th a

nd

GDP

per c

apita

•tra

de o

penn

ess

of h

ost c

ount

ry•

pres

ence

of P

TIAs

•ho

st c

ount

ry p

oliti

cal r

isk

•po

litic

al c

onst

rain

ts a

nd le

vel o

f de

moc

racy

•en

d of

the

cold

war

•di

stan

ce b

etw

een

the

US a

nd h

ost

coun

try

Fixe

d ef

fect

est

imat

ion

Ther

e is

a p

ositi

ve im

pact

of r

atifi

ed

BITs

on

FDI.

The

impa

ct o

f rat

ified

BIT

s is

not

onl

y st

atis

tical

ly si

gnifi

cant

(at l

east

at 9

0%

confi

denc

e) b

ut a

lso

subs

tant

ive: a

jo

intly

ratifi

ed B

IT in

crea

ses

US fo

reig

n in

vest

men

t in

the

host

cou

ntry

from

0.

07 to

0.2

4% o

f the

dom

estic

GDP

. In

cont

rast

, the

impa

ct o

f sig

ned

BITs

is

posi

tive

but n

ot s

tatis

tical

ly si

gnifi

cant

.

The

empi

rical

ana

lysis

pro

ves

that

onl

y ra

tified

BIT

s ha

ve

stat

istic

ally

sign

ifica

nt e

ffect

s on

FDI

as

they

func

tion

as a

co

stly

sign

al o

f a p

ro-in

vest

men

t clim

ate

and

of a

cre

d-ib

le c

omm

itmen

t fro

m th

e ho

st c

ount

ry to

the

prot

ectio

n of

FDI

.

Oh a

nd F

rati-

anni

(201

0)Bi

late

ral F

DI fl

ows

1980

-200

514

8 co

untri

es•

Exis

tenc

e of

a B

IT•

tota

l num

ber o

f BIT

s by

th

e tw

o co

untri

es

•GD

P•

popu

latio

n, c

omm

on la

nd b

orde

r, di

stan

ce, c

omm

on la

ngua

ge, c

omm

on

colo

nial

her

itage

, com

mon

cur

renc

y, m

embe

rs o

f sam

e RT

A•

dem

ogra

phic

, ins

titut

iona

l and

cul

tura

l va

riabl

es

A gr

avity

mod

el e

sti-

mat

ed w

ith fi

xed

effe

ct

and

Pois

son

quas

i m

axim

umlik

elih

ood

estim

ator

(Q

MLE

)

Ther

e is

a s

tatis

tical

ly si

gnifi

cant

and

po

sitiv

e im

pact

of B

ITs

on F

DI.

The

stoc

k of

BIT

s is

sub

ject

to d

imin

ish-

ing

retu

rns

mea

sure

d in

term

s of

FDI

flo

ws.

TNCs

find

mor

e va

lue

in in

vest

ing

whe

re a

bila

tera

l tre

aty

is in

pla

ce.

17

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Pint

o, P

into

an

d St

ier-

Mo-

ses

(201

0)

Bila

tera

l flow

s fro

m th

e Un

ited

Stat

es to

ratif

ying

econ

omie

s

1970

-200

6Ho

me:

Uni

ted

Stat

esHo

st: 6

7 co

untri

es

that

hav

e ra

tified

BI

Ts w

ith th

e Un

ited

Stat

es

•Ra

tified

BIT

s •

GDP

per c

apita

•tra

de o

penn

ess

•ex

chan

ge ra

te•

mar

ket s

ize (p

opul

atio

n)

•as

par

t of a

gra

vity

mod

el d

umm

ies

for a

com

mon

bor

der a

nd la

ngua

ge

betw

een

the

Unite

d St

ates

and

the

host

cou

ntry

•

tem

pora

l dum

mie

s of

four

, five

and

te

n ye

ars

inte

rval

s si

nce

sign

ing

a BI

T •

lagg

ed d

epen

dent

var

iabl

e

A gr

avity

mod

el

Ratifi

ed B

ITs

have

sig

nific

ant a

nd s

izabl

e ef

fect

on

FDI fl

ows.

Add

ing

tem

pora

l du

mm

ies,

how

ever

, red

uces

the

size

and

si

gnifi

canc

e le

vel o

f BIT

s on

out

war

d US

in

vest

men

t.

The

effe

ct o

f BIT

s on

FDI

is n

ot s

tabl

e. T

he e

ffect

dis

-si

pate

s ov

er ti

me.

BIT

s an

d go

od d

omes

tic in

stitu

tions

are

co

mpl

emen

tary

in a

ttrac

ting

FDI.

Desb

orde

s an

d Vi

card

(200

9)Bi

late

ral F

DI s

tock

19

91-2

000

Hom

e: 3

0 OE

CD

coun

tries

Host

: 32

non-

OECD

co

untri

es

•En

try in

to fo

rce

of B

ITs

•M

arke

t size

(GDP

)•

GDP

per c

apita

•

qual

ity o

f dom

estic

inst

itutio

ns

•qu

ality

of i

nter

stat

e re

latio

ns

A gr

avity

mod

el u

sing

Po

isso

n qu

asi m

axi-

mum

like

lihoo

d es

timat

or

(QM

LE)

BITs

hav

e a

grea

ter e

ffect

whe

n im

ple-

men

ted

betw

een

coun

tries

with

pol

itica

l te

nsio

ns w

hile

they

hav

e no

sig

nific

ant

effe

ct b

etw

een

frien

dly

coun

tries

.

BITs

and

goo

d do

mes

tic in

stitu

tions

are

com

plem

enta

ry

in a

ttrac

ting

FDI.

Kern

er (2

009)

Bila

tera

l (lo

g) F

DI

at c

onst

ant 2

000

dolla

rs

1982

-200

1Ho

me:

OEC

D co

un-

tries

Host

: 127

dev

elop

-in

g co

untri

es

•Ra

tified

BIT

s an

d ot

her

IIAs

with

oth

er O

ECD

coun

tries

•M

arke

t size

(GDP

)•

savin

g ra

tes

•tra

de o

penn

ess

•pr

esen

ce o

f a P

TIA

•de

moc

racy

A gr

avity

mod

el w

ith

inst

rum

enta

l var

iabl

e to

tack

le e

ndog

enei

ty

prob

lem

.

Whe

n es

timat

ing

with

inst

rum

enta

l va

riabl

es B

ITs

are

pos

itive

ly an

d si

gnifi

-ca

ntly

corre

late

d w

ith F

DI in

flow

s

BITs

attr

act s

igni

fican

t am

ount

s of

inve

stm

ent.

BITs

attr

act t

his

inve

stm

ent f

rom

pro

tect

ed a

nd u

npro

-te

cted

inve

stor

s.

Aisb

ett (

2009

)Bi

late

ral (

log)

in

flow

s of

FDI

1980

-199

928

dev

elop

ing

coun

tries

; 29

OECD

co

untri

es Le

ss th

an 6

72

obse

rvat

ions

per

ye

ar.

•La

gged

BIT

ratifi

catio

n•

Host

and

hom

e co

untry

GDP

•po

pula

tion

•sh

are

of tr

ade

in G

DP•

the

skill

gap

betw

een

host

and

hom

e co

untry

Fixe

d ef

fect

s es

timat

ion

BITs

are

pos

itive

ly an

d si

gnifi

cant

ly co

rrela

ted

with

FDI

inflo

ws.

But

this

do

es n

ot im

ply

that

BIT

s ca

use

larg

e FD

I inc

reas

e (o

f up

to 5

0%) b

ecau

se o

f en

doge

neity

of B

ITs.

The

lack

of e

viden

ce o

n th

e im

pact

of B

ITs

on F

DI is

im

porta

nt in

form

atio

n fo

r cou

ntrie

s w

eigh

ing

the

cost

and

be

nefit

s of

beg

inni

ng o

r exp

andi

ng B

ITs

prog

ram

mes

.

Bang

a (2

008)

FDI i

nflow

s ba

sed

on a

ppro

ved

FDI

from

dev

elop

ed

and

deve

lopi

ng

coun

tries

1980

-81

to 1

999-

2000

; 19

86-8

7 to

19

96-9

7 fo

r pan

el

data

for 1

0 ho

st c

oun-

tries

15 h

ost d

evel

op-

ing

coun

tries

from

So

uth,

Eas

t and

So

uth-

East

Asi

a; a

ll ho

me

deve

lope

d an

d de

velo

ping

co

untri

es

•FD

I pol

icie

s su

ch a

s FD

I lib

eral

izatio

n, in

cent

ives,

pr

ofit t

rans

fer a

nd ta

riffs

•to

tal n

umbe

r of B

ITs

and

dum

mie

s fo

r APE

C an

d AS

EAN

inve

stm

ent

agre

emen

ts

•GD

P an

d GD

P gr

owth

•w

ages

, edu

catio

n an

d la

bour

pro

-du

ctivi

ty•

cost

of c

apita

l•

exch

ange

rate

•in

frast

ruct

ure

•de

bt a

nd b

udge

t defi

cit

Rand

om e

ffect

s es

ti-m

atio

nFD

I pol

icie

s ar

e an

impo

rtant

det

er-

min

ant o

f FDI

inflo

ws,

esp

ecia

lly th

e re

mov

al o

f res

trict

ions

. BIT

s pl

ay a

n im

porta

nt ro

le in

stim

ulat

ing

inflo

ws,

es-

peci

ally

BITs

with

dev

elop

ed c

ount

ries.

18

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Pein

hard

t and

Al

lee

(200

8)Bi

late

ral fl

ows

from

the

Unite

d St

ates

1966

-200

717

8 ho

st c

ount

ries

•En

try in

to fo

rce

of B

ITs,

Tr

ade

and

Inve

stm

ent

Fram

ewor

k Ag

reem

ents

(T

IFAs

) and

PTI

As

•Re

al G

DP•

GDP

per c

apita

•US

trad

e ag

reem

ents

•m

embe

rshi

p in

WTO

, com

mitm

ents

to

prop

erty

righ

ts

Fixe

d ef

fect

est

imat

ion

of lo

g-lin

ear m

odel

Depe

ndin

g on

the

mod

el, P

TIAs

hav

e st

rong

er e

ffect

on

FDI t

han

TIFA

s an

d BI

Ts.

Ther

e is

littl

e ev

iden

ce th

at U

S tra

de a

nd in

vest

men

t ag

reem

ents

hav

e ef

fect

s on

FDI

.

Sieg

man

n (2

008)

Bila

tera

l FDI

flow

s19

80-2

004

OECD

cou

ntrie

s as

hom

e an

d 62

de

velo

ping

cou

ntrie

s as

reci

pien

ts

•Ex

iste

nce

of a

rati-

fied

BIT

•Gr

avity

mod

el:

•GD

P ho

me,

•

GDP

host

, •

popu

latio

n ho

me,

•

popu

latio

n ho

st,

•ov

eral

l tra

de h

ome,

•

trade

hos

t, •

skill

diffe

renc

e

A gr

avity

mod

el; i

n ad

ditio

n a

know

ledg

e-ca

pita

l mod

el is

test

ed

Ther

e is

an

empi

rical

ly va

lid, p

ositi

ve

influ

ence

of B

ITs

and

DTTs

on

FDI.

The

effe

ct a

ppea

rs to

rang

e be

twee

n 30

per

ce

nt a

nd 4

0 pe

r cen

t for

BIT

s.

The

effe

ct o

f enf

orce

d BI

Ts is

dis

tinct

and

stro

ng, b

ut th

e si

gnat

ure

of a

BIT

doe

s no

t app

ear t

o ha

ve a

sig

nific

ant

impa

ct o

n FD

I flow

s.

A co

mpl

emen

tary

rela

tions

hip

exis

ts b

etw

een

BITs

and

in

stitu

tiona

l qua

lity:

BIT

s ar

e m

ore

effe

ctive

if th

ey a

re

enfo

rced

in a

n en

viron

men

t of h

igh

polit

ical

sta

bilit

y an

d an

effe

ctive

lega

l sys

tem

.

The

effe

ctive

ness

of B

ITs

appe

ars

to b

e gr

eate

r, if

both

pa

rtner

cou

ntrie

s ar

e ra

tifyin

g m

embe

rs o

f the

ICSI

D Co

nven

tion.

The

mac

roec

onom

ic e

nviro

nmen

t of a

hos

t cou

ntry

ap

pear

s to

hav

e no

influ

ence

on

a BI

T’s

effe

ctive

ness

. Li

kew

ise,

exc

hang

e ra

te v

olat

ility,

as w

ell a

s cu

rrenc

y cr

ises

app

ear t

o ha

ve n

o si

gnifi

cant

impa

ct.

Lesh

er a

nd

Miro

udot

(2

007)

Bila

tera

l FDI

flow

s19

90-2

004

182

•Au

thor

-cre

ated

BIT

in

dex

•BI

T du

mm

y va

riabl

es

•Di

stan

ce•

bord

er•

exch

ange

rate

s•

GDP

A gr

avity

mod

el fo

r de

term

inan

ts o

f FDI

Inve

stm

ent p

rovis

ions

in R

TAs

are

posi

-tiv

ely

asso

ciat

ed w

ith in

vest

men

t flow

s.

Ther

e is

an

insi

gnifi

cant

effe

ct o

f BIT

s on

inve

stm

ent fl

ows.

Subs

tant

ive in

vest

men

t pro

visio

ns in

RTA

s im

pact

trad

e an

d FD

I flow

s m

ore

prof

ound

ly or

the

com

bina

tion

of s

ub-

stan

tive

inve

stm

ent r

ules

and

pro

visio

ns li

bera

lisin

g ot

her

parts

of t

he e

cono

my

join

tly im

pact

trad

e an

d in

vest

men

t m

ore

sign

ifica

ntly.

Yack

ee (2

007)

As in

Neu

may

er

and

Spes

s (2

005)

, bu

t add

ing

a co

n-st

ant s

tart

to F

DI

shar

e va

riabl

e

1984

-200

3Ho

st d

evel

opin

g co

untri

es a

nd 1

8 ho

me

capi

tal e

x-po

rting

cou

ntrie

s

•A

wei

ghte

d co

unt o

f BI

Ts s

igne

d by

cap

ital

impo

rting

cou

ntrie

s w

ith

18 c

apita

l-exp

ortin

g co

untri

es•

in a

dditi

on to

BIT

s, F

CN

treat

ies

and

FTAs

with

in

vest

men

t pro

visio

ns

•As

in N

eum

ayer

and

Spe

ss (2

005)

, w

ith tr

ade

open

ness

repl

acin

g th

e nu

mbe

r of F

TAs

Sam

e m

etho

dolo

gy a

s in

Neu

may

er a

nd S

pess

(2

005)

, with

som

e ch

ange

s, e

.g. e

xclu

ding

ye

ar d

umm

ies

The

case

for B

ITs

is fa

r wea

ker t

han

sugg

este

d by

Neu

may

er a

nd S

pess

(2

005)

.

Smal

l cha

nges

in m

etho

dolo

gy a

nd

mod

el s

peci

ficat

ions

mak

e th

e BI

Ts e

f-fe

ct o

n FD

I lar

gely

or e

ntire

ly di

sapp

ear.

The

inst

itutio

nal q

ualit

y te

st s

how

s an

op

posi

te c

ondi

tiona

l rel

atio

nshi

p th

an

that

foun

d by

Neu

may

er a

nd S

pess

(2

005)

.

19

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Egge

r and

M

erlo

(200

7)Bi

late

ral s

tock

s of

ou

twar

d FD

I19

80-2

001

24 h

ome

and

28

host

cou

ntrie

s;

22 o

f the

28

host

co

untri

es a

re O

ECD

coun

tries

•Du

mm

y va

riabl

e fo

r BIT

ra

tifica

tion

(or c

oncl

u-si

on)

•On

ce-la

gged

FDI

sto

cks

•jo

int s

ize o

f hom

e an

d ho

st c

ount

ry

mar

kets

in te

rms

of G

DP•

hom

e-to

-hos

t cou

ntry

rela

tive

GDP

•ho

me-

to-h

ost c

ount

ry s

kille

d-la

bour

en

dow

men

t rat

io

Gene

ralis

ed M

etho

d of

Mom

ents

(GM

M)

estim

atio

n

The

varia

ble

for B

IT ra

tifica

tion

has

a po

sitiv

e an

d si

gnifi

cant

impa

ct o

n ou

twar

d FD

I sto

cks.

The

shor

t-ru

n im

pact

of B

ITs

is s

mal

ler

than

the

long

-run

impa

ct.

Ther

e is

a s

ubst

antia

l diff

eren

ce b

etw

een

the

exis

ting

posi

tive

shor

t-ru

n an

d lo

ng-r

un im

pact

of B

ITs

on F

DI.

Henc

e th

ere

is a

nee

d to

take

the

dyna

mic

nat

ure

of F

DI

mor

e in

to a

ccou

nt.

Tobi

n an

d Ro

se-A

cker

-m

an (2

006)

Tota

l FDI

inflo

ws

into

dev

elop

-in

g co

untri

es in

co

nsta

nt 2

000

dolla

rs;

OECD

out

flow

s to

de

velo

ping

cou

n-tri

es in

con

stan

t do

llars

;fiv

e-ye

ar a

vera

ges

1980

-200

313

7 de

velo

ping

co

untri

es (i

ncre

ased

fro

m 4

0 de

velo

ping

co

untri

es w

ith a

ll da

ta, u

sing

bes

t da

ta p

redi

ctio

ns a

nd

othe

r tec

hniq

ues)

•To

tal n

umbe

r of B

ITs

•to

tal n

umbe

r of B

ITs

with

dev

elop

ing

coun

-tri

es•

sign

ed B

ITs

•w

eigh

ted

and

un-

wei

ghte

d BI

Ts in

dex

by

the

size

of t

he h

ome

OECD

cou

ntry

; int

er-

actio

n be

twee

n ho

st

coun

try B

ITs

and

tota

l nu

mbe

r of B

ITs

in th

e w

orld

•Po

litic

al ri

sk•

GDP

per c

apita

•po

pula

tion

•GD

P gr

owth

•na

tura

l res

ourc

e en

dow

men

ts•

trade

/GDP

Fixe

d-ef

fect

s es

tima-

tion

The

num

ber o

f BIT

s w

ith h

igh

inco

me

coun

tries

has

a p

ositi

ve a

nd s

igni

fican

t ef

fect

on

FDI i

nflow

s.

Mor

e w

orld

wid

e BI

Ts re

duce

the

mar

-gi

nal b

enefi

t of a

n ex

tra B

IT to

a h

ost

coun

try.

As e

ach

extra

BIT

has

dec

reas

ing

bene

fits

in te

rms

of

stim

ulat

ing

FDI i

nflow

s, h

ost c

ount

ries

may

be

less

eag

er

to s

ign

BITs

ove

r tim

e.

Galla

gher

and

Bi

rch

(200

6)To

tal F

DI in

flow

s;

FDI i

nflow

s fro

m

the

US

1980

-200

324

hos

t cou

ntrie

s fro

m L

atin

Am

eric

a •

Tota

l num

ber o

f BIT

s•

BITs

with

the

US•

Infla

tion

•GD

P•

tota

l exp

orts

or e

xpor

ts/G

DP•

liter

acy

rate

•GD

P pe

r cap

ita•

GDP

grow

th•

num

ber o

f priv

atiza

tions

Fixe

d-ef

fect

s es

tima-

tion

Neith

er th

e to

tal n

umbe

r of B

ITs

nor

BITs

with

the

US h

ave

an in

depe

nden

t an

d po

sitiv

e ef

fect

on

tota

l FDI

inflo

ws

or

inflo

ws

from

the

US.

It m

ay n

ot b

e w

orth

whi

le to

car

ry th

e co

sts

of B

ITs

such

as

lifti

ng p

erfo

rman

ce re

quire

men

ts a

nd a

pplyi

ng b

road

ex

prop

riatio

n ru

les.

Gros

se a

nd

Trev

ino

(200

5)An

nual

FDI

in-

flow

s in

to a

hos

t co

untry

1990

-199

913

hos

t cou

ntrie

s fro

m C

entra

l and

Ea

ster

n Eu

rope

•A

tota

l num

ber o

f BIT

s co

nclu

ded

by a

hos

t co

untry

•th

e im

pact

of B

ITs

was

ex

amin

ed in

the

cont

ext

of o

ther

inst

itutio

nal

varia

bles

rela

ted

to

corru

ptio

n, re

gula

tions

on

FDI

and

ent

erpr

ises

, pr

ivatiz

atio

n an

d po

liti-

cal r

isk

•In

flatio

n•

curre

ncy

valu

atio

n an

d m

arke

t size

Stan

dard

mul

tiple

re

gres

sion

BITs

tend

to s

timul

ate

inw

ard

FDI,

toge

ther

with

the

degr

ee o

f ent

erpr

ise

refo

rm a

nd re

patri

atio

n ru

les.

The

findi

ng s

houl

d be

of i

nter

est t

o in

tern

atio

nal o

rgan

iza-

tions

and

to h

ost d

evel

opin

g co

untri

es.

20

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Neum

ayer

and

Sp

ess

(200

5)FD

I infl

ows

into

a

host

cou

ntry

in

cons

tant

199

6 US

$; s

hare

of a

ho

st c

ount

ry’s

to

tal i

nflow

s of

de

velo

ping

cou

n-tri

es

1970

-20

01 a

nd

1984

-200

1 fo

r som

e va

riabl

es

119

deve

lopi

ng

coun

tries

; OEC

D co

untri

es

Num

ber o

f BIT

s w

ith

OECD

cou

ntrie

s w

eigh

ted

by a

sha

re o

f a h

ome

coun

try in

wor

ld o

utw

ard

FDI fl

ows

•Lo

g of

per

cap

ita G

DP a

nd p

opul

a-tio

n si

ze•

GDP

grow

th ra

te•

WTO

mem

bers

hip

•nu

mbe

r of B

ITs

with

hom

e co

untri

es•

infla

tion

rate

•na

tura

l res

ourc

e in

tens

ity•

polit

ical

sta

bilit

y•

inst

itutio

nal q

ualit

y•

inve

stm

ent r

isk

inde

x•

trade

ope

nnes

s an

d se

cond

ary

enro

l-m

ent i

n se

nsiti

vity

anal

ysis

Rand

om a

nd fi

xed-

effe

cts

estim

atio

nsA

posi

tive

effe

ct o

f BIT

s w

ith d

evel

oped

co

untri

es o

n FD

I was

foun

d, w

hich

is

cons

iste

nt a

nd ro

bust

acr

oss

vario

us

mod

el s

peci

ficat

ions

. Som

etim

es th

e ef

fect

dep

ends

on

inst

itutio

nal q

ualit

y.

The

unde

rtaki

ng o

f the

obl

igat

ions

con

tain

ed in

BIT

s by

de

velo

ping

cou

ntrie

s do

es h

ave

a de

sire

d pa

yoff

of h

ighe

r FD

I infl

ows.

Dev

elop

ing

coun

tries

that

sig

n m

ore

BITs

with

de

velo

ped

coun

tries

rece

ive m

ore

FDI.

But i

t is

impo

ssib

le

to te

ll if

bene

fits

from

incr

ease

d FD

I infl

ows

are

high

er

than

cos

ts o

f BIT

s fo

r dev

elop

ing

coun

tries

.

Sala

cuse

and

Su

llivan

(200

5)1.

Tot

al F

DI in

-flo

ws

(% c

hang

es)

2. B

ilate

ral F

DI

flow

s fro

m th

e US

1998

, 199

9 an

d 20

00

1991

-200

0

Mor

e th

an 1

00

deve

lopi

ng c

ount

ries

31 d

evel

opin

g co

untri

es; U

SA30

0 ob

serv

atio

ns

•A

US B

IT: a

tota

l num

ber

of B

ITs

with

oth

er O

ECD

coun

tries

; a to

tal w

ith

deve

lopi

ng c

ount

ries

• A

BIT

with

the

US;

num

ber o

f oth

er O

ECD

BITs

• H

ost c

ount

ry G

DP•

GDP

per c

apita

•re

al e

ffect

ive e

xcha

nge

rate

•po

pula

tion

•ru

le o

f law

As a

bove

plu

s to

tal F

DI in

flow

s.

Mul

tivar

iate

OLS

and

cr

oss-

sect

iona

l reg

res-

sion

Fixe

d-ef

fect

s es

tima-

tion

US B

ITs

have

a la

rge,

pos

itive

and

sig

-ni

fican

t ass

ocia

tion

with

a h

ost c

ount

ry’s

ov

eral

l FDI

inflo

ws.

The

impa

ct o

f oth

er

OECD

BIT

s is

wea

ker.

Ther

e is

stro

ng e

viden

ce th

at B

ITs

have

atta

ined

to a

si

gnifi

cant

ext

ent t

heir

stat

ed g

oal o

f pro

mot

ing

FDI.

It is

be

tter t

o si

gn B

ITs

with

hig

her p

rote

ctio

n st

anda

rds

(like

th

ose

of U

S BI

Ts).

Büth

e an

d M

ilner

(200

4)An

nual

inflo

ws

of F

DI in

to h

ost

coun

tries

1970

-200

0Up

to 1

22 h

ost

deve

lopi

ng c

ount

ries

with

a p

opul

atio

n ov

er 1

milli

on

•A

tota

l num

ber o

f si

gned

cum

ulat

ive B

ITs

•M

arke

t size

•ec

onom

ic d

evel

opm

ent

•ec

onom

ic g

row

th•

trade

ope

nnes

s•

dom

estic

pol

itica

l con

stra

ints

and

po

litic

al in

stab

ility

Fixe

d ef

fect

ana

lyses

; ad

ditio

nal t

ests

for

hete

rosk

edas

ticity

and

au

toco

rrela

tion

A st

atis

tical

ly an

d su

bsta

ntia

lly s

ig-

nific

ant c

orre

latio

n be

twee

n BI

Ts a

nd

subs

eque

nt in

war

d FD

I int

o de

velo

ping

co

untri

es w

as fo

und.

Each

dev

elop

ing

coun

try h

as to

wei

gh th

e co

sts

of B

ITs

agai

nst t

he b

enefi

ts o

f inc

reas

ed F

DI a

nd p

ossi

bly

othe

r be

nefit

s.

Egge

r and

Pf

affe

rmay

r (2

004)

Bila

tera

l sto

cks

of o

utw

ard

FDI i

n co

nsta

nt 1

995

US

dolla

rs

1982

-199

719

hom

e OE

CD

coun

tries

(old

and

ne

w) a

nd 5

7 ho

st

coun

tries

(inc

ludi

ng

27 O

ECD

coun

tries

)

•Si

gned

BIT

s an

d ra

tified

BI

Ts b

etw

een

coun

tries

in

the

sam

ple

•Co

untry

size

(GDP

)•

fact

or e

ndow

men

ts (t

ertia

ry

enro

lmen

t)•

trade

and

FDI

fric

tion

and

inte

ract

ion

term

s

In s

ome

spec

ifica

tions

in a

dditi

on (o

r in

stea

d):

•Eu

rope

an U

nion

(EU)

and

Nor

th

Amer

ican

Fre

e Tr

ade

Agre

emen

t (N

AFTA

) mem

bers

hip

•re

al G

DP p

er c

apita

or s

econ

dary

sc

hool

enr

olm

ent (

for f

acto

r end

ow-

men

ts)

Fixe

d-ef

fect

s es

tima-

tion

A po

sitiv

e im

pact

of r

atifi

ed B

ITs

on F

DI

on b

ilate

ral s

tock

s of

FDI

was

est

ab-

lishe

d.

In m

ost o

f the

cas

es th

e po

sitiv

e im

pact

of

eve

n on

ly si

gnin

g BI

Ts o

n ou

twar

d FD

I ex

iste

d, a

lthou

gh a

t a lo

wer

sig

nific

ance

le

vel.

Impl

emen

ted

BITs

exe

rt a

posi

tive

and

sign

ifica

nt e

ffect

on

real

sto

cks

of o

utw

ard

FDI.

21

Sour

ceDe

pend

ent

varia

ble

Perio

dHo

st a

nd h

ome

coun

trie

sEx

plan

ator

y va

riabl

e re

-la

ted

to B

ITs

and

PTIA

sCo

ntro

l var

iabl

esEc

onom

etric

met

hod

Econ

omet

ric re

sults

Conc

lusi

ons

Tobi

n an

d Ro

se-A

cker

-m

an (2

003)

1. T

otal

FDI

in

flow

s, 5

-yea

r av

erag

es; s

hare

of

wor

ld in

flow

s

2. F

DI in

flow

s fro

m th

e US

1975

-200

0

1980

-200

0

45 d

evel

opin

g co

untri

es

48 d

evel

opin

g co

untri

es

•To

tal n

umbe

r of B

ITs

•nu

mbe

r of B

ITs

with

hi

gh-

and

low

-inco

me

coun

tries

•A

BIT

with

the

US

•Lo

g of

GDP

per

cap

ita•

popu

latio

n•

fuel

s an

d or

es e

xpor

ts•

blac

k m

arke

t rat

e of

exc

hang

e pr

emia

•ag

greg

ate

polit

ical

-ris

k•

host

cou

ntry

dis

tanc

e fro

m e

quat

or

Sam

e sp

ecifi

catio

ns a

s ab

ove

plus

di

stan

ce o

f hos

t cou

ntry

from

the

US;

exch

ange

rate

sta

bilit

y; s

kill

diffe

renc

es

Fixe

d-ef

fect

s es

tima-

tion;

on

som

e ca

ses

rand

om e

ffect

s es

tima-

tion

is c

onsi

dere

d Fi

xed-

effe

cts

estim

a-tio

n (in

som

e ca

ses

pool

ed e

stim

atio

n)

BITs

app

ear t

o ha

ve li

ttle

impa

ct o

n FD

I. Ne

gativ

e ef

fect

s at

the

high

leve

l of r

isk.

Po

sitiv

e ef

fect

at t

he lo

w le

vel o

f ris

k.

The

maj

ority

of d

evel

opin

g co

untri

es a

re

in th

e hi

gh ri

sk c

ateg

ory.

Littl

e re

latio

nshi

p be

twee

n a

US B

IT

and

inflo

ws

from

the

US. W

here

ther

e is

re

latio

nshi

p it

is w

eakl

y ne

gativ

e.

Hallw

ard-

Drie

-m

eier

(200

3)Bi

late

ral fl

ows

of

FDI;

inflo

ws/

GDP;

sh

are

of h

ome

coun

try o

utflo

ws

1980

-200

031

hos

t dev

elop

ing

coun

tries

; 20

OECD

co

untri

es

537

pairs

of c

oun-

tries

•Co

nclu

sion

of a

BIT

- The

size

of t

he h

ost a

nd h

ome

coun

try-

infla

tion

- tra

de/G

DP-

skills

gap

- co

mpo

nent

s of

inst

itutio

nal q

ualit

y fro

m IC

RG (l

egal

sys

tem

and

cor

rupt

ion)

In a

dditi

on, t

rans

ition

to a

mar

ket e

cono

-m

y an

d th

e co

nclu

sion

of N

AFTA

.

Fixe

d-ef

fect

s es

tima-

tions

No s

tatis

tical

ly si

gnifi

cant

effe

ct o

f BIT

s on

FDI

inflo

ws

was

foun

d.BI

Ts a

re c

ompl

emen

ts to

goo

d in

stitu

tiona

l qua

lity

rath

er

than

sub

stitu

tes.

UNCT

AD

(199

8b)

1. B

ilate

ral F

DI

flow

s; s

hare

of

host

cou

ntry

in

hom

e co

untry

’s

tota

l out

flow

s;

shar

e of

hom

e co

untry

in to

tal

host

cou

ntry

’s F

DI

inflo

ws

2. T

otal

FDI

in

flow

s in

to a

ho

st c

ount

ry; F

DI

stoc

ks; F

DI/G

DP

1971

-199

4

1995

72 h

ost d

evel

opin

g co

untri

es; 1

4 OE

CD

coun

tries

133

host

dev

elop

ing

coun

tries

•Co

nclu

sion

of a

BIT

•A

tota

l num

ber o

f BIT

s co

nclu

ded

by a

hos

t co

untry

•GD

P of

hos

t cou

ntry

•po

pula

tion

•GD

P•

popu

latio

n•

dom

estic

inve

stm

ent o

f hos

t cou

ntrie

s

Cros

s-se

ctio

nal s

tep-

wis

e re

gres

sion

s

BITs

cou

ld c

ause

sm

all i

ncre

ase

of F

DI

from

a h

ome

partn

er c

ount

ry. B

ut re

sults

ar

e no

t rob

ust.

Smal

l red

irect

ion

of F

DI

to B

IT p

artn

ers.

BITs

foun

d to

hav

e a

posi

tive

and

sta-

tistic

ally

sign

ifica

nt e

ffect

in th

ree

out o

f ni

ne re

gres

sion

s.

BITs

app

ear t

o pl

ay a

min

or a

nd s

econ

dary

role

in in

flu-

enci

ng F

DI fl

ows.

So

urc

e:

UN

CTA

D.

22

For the latest investment trends and policy developments, including International Investment Agreements (IIAs), visit the website of the UNCTAD Investment and Enterprise Division: www.unctad.org/diae and www.unctad.org/iia

For further information, please contact Mr. James X. Zhan Director Investment and Enterprise Division – UNCTAD

Tel.: 00 41 22 917 57 60 Fax: 00 41 22 917 04 98

Follow us on @unctadwif

http://unctad-worldinvestmentforum.org

Join UNCTAD’s World Investment Forum 2014 for acomprehensive discussion between investment stakeholders

on the best reform options for ISDS and the IIA regime.

U N I T E D N AT I O N S C O N F E R E N C E O N T R A D E A N D D E V E L O P M E N T

WORLD INVESTMENT

REPORT2014INVESTING IN THE SDGs: AN ACTION PLAN World Investment Report 2014:

Investing in the SDGs: An Action Planwww.unctad.org/WIR