The (Political) Economics of Bilateral Investment Treaties—The Unique Trajectory of Brazil

Abstract

Brazil, after signing several traditional Bilateral Investment Treaties without ratifying them, recently shifted towards a different type of bilateral investment agreement, i.e., Investment Cooperation and Facilitation Agreements. Two claims have been made in the literature regarding the transition from traditional Bilateral Investment Treaties to Investment Cooperation and Facilitation Agreements—Claim #1: The non-ratification of the traditional BITs has not harmed Foreign Direct Investment into Brazil, a claim which puts into question the purpose of Bilateral Investment Treaties. Claim #2: While Investment Cooperation and Facilitation Agreements avoid some of the problems of traditional Bilateral Investment Treaties, on balance they are less effective than traditional Bilateral Investment Treaties would have been. We examine the two claims from an empirical economic point of view. We build on the literature about Brazil’s position vis-à-vis Bilateral Investment Treaties, which must be viewed by an amalgamation of (i) a historical legacy; (ii) domestic initiatives, and (iii) a particular U-turn in the political debate. Using empirical evidence on Foreign Direct Investment effects of Bilateral Investment Treaties, the following conclusions emerge: With regard to claim #1, empirical evidence in general as well as specific to Brazil suggests that Brazil has forgone Foreign Direct Investment by not ratifying traditional Bilateral Investment Treaties. Concerning claim #2, while Investment Cooperation and Facilitation Agreements include alternative dispute settlement mechanisms, which aim at a better compliance of states with the Investment Cooperation and Facilitation Agreements’ rules, rather than the compensation of foreign investors, the lower stringency of the State–State dispute settlement mechanism compared to Investor–State dispute settlement mechanism makes Investment Cooperation and Facilitation Agreements less effective. Yet, this weakening effect must be weighed against the effects on Foreign Direct Investment from innovative clauses in Investment Cooperation and Facilitation Agreements, which are absent in many traditional Bilateral Investment Treaties.

1. Introduction

Foreign Direct Investment (FDI) is considered as an important part of a country’s development strategy.1 Developing countries’ governments regularly compete for FDI, often with neighboring countries of similar levels of development. One possible route taken by countries is to conclude, i.e., sign and ratify, International Investment Agreements (IIAs). Dolzer and Schreuer (2008) (p. 9) even stress that “investment treaties are today seen as admission tickets to international investment markets. Their limiting impact on the sovereignty of the host state, controversial as it may be in the individual case, is in this sense a necessary corollary to the objective of creating an investment-friendly climate”.

Investment treaties are based on the presumption of the existence of a hold-up problem (Hart 1995) in international investment. Over time, more and more developing countries have concluded IIAs, which fall into Bilateral Investment Treaties (BITs), treaties with investment provisions and investment-related instruments.2 Clearly, IIAs are just one puzzle stone in the puzzle of location factors that matter in an investment decision.3

Despite subtle differences, Elkins et al. (2006, p. 7) emphasize the fact that in the past, there has been one dominant treaty model, mainly reflecting the interests of developed nations. Developing countries could opt to take it or to leave it. A similar conclusion has been reached by Alschner and Skougarevskiy (2016, p. 562) who find that “wealthier states achieve considerably more consistent IIA networks than their poorer counterparts, which lends support to the argument that developed countries are the IIA system’s rule-makers, while developing countries tend to be its rule-takers”. While this is undisputed, not all developing countries have been rule-takers at all times.

A case in question is Brazil, which undoubtedly enacted trade policies (e.g., negotiating the trade agreement between the European Union and Mercosur, Baur et al. 2023) and other economic policies, like privatization programs, in line with the Washington consensus and conducive to FDI. In Brazil, the organization of companies and their interaction with the state are relatively weak, which resulted in policies less conducive to domestic business. Instead, a policy oriented towards financial stability aiming at attracting foreign investors has been conducted (Pedersen 2009, p. 472).

Brazil followed a special path with respect to FDI policies. “It is widely acknowledged that Brazil has been the most reticent to conclude BITs as a result of national concerns over the effect of such agreements on limiting developmental policy space (Barbosa 2003).” (quoted from Dixon and Haslam 2016, p. 1091). Indeed, Brazil never ratified any traditional BITs, but it has recently ratified a new type of BIT, namely the Investment Cooperation and Facilitation Agreements (CFIAs), which deviate from traditional BITs substantively (i.e., regarding the clauses included) and procedurally (i.e., regarding the enforcement mechanism included).

The current (traditional) international investment policy regime, notably BITs, is confronted with substantial criticism. For example, Tienhaara et al. (2023, p. 1197) state that the enforcement mechanism included in traditional BITs “likely obstruct(s) a just transition by chilling supply-side climate measures and diverting public funds away from climate change mitigation and adaptation efforts”. Brazil’s CFIAs could potentially act as a role model for a new generation of BITs, which avoids the pitfalls of traditional BITs (like an asserted chilling effect on public regulation) but at the same time also acts as a promotor of FDI.

With this paper, we aim at providing a systematic and structured review of the likely effects of Brazil’s old and new BIT policies on FDI. Our analysis adds new insights to the discussion of “FDI, BITs, and the Brazilian way” as hitherto the literature on Brazil’s CFIAs mainly focused on legal issues and has largely neglected economic and political science issues. The findings of this review will help policy makers in their attempt to form a new institutional set-up for international investment.

To achieve this goal, we have surveyed the legal, economic, and political science literature dealing with effects of BITs on FDI in general and with the absence of traditional BITs in Brazil in particular. To put structure in the survey, we use two claims made by scholars regarding the effectiveness of Brazil’s BIT policies for FDI (see Section 2). We surveyed the theoretical and empirical literature to provide our answer regarding the empirical validity of the two claims made. Thereby, we focused only on studies available in the English language and we left out the vast legal literature written in the Portuguese language. To avoid deviations from the original content of and conveyed implications by the surveyed papers, we frequently use direct citations (in “ ”). Finally, throughout this review, we focused on inward FDI, while it is undisputed that BITs also affect outward FDI and indeed a central motivation of Brazil to enter into CFIA negotiations is to stimulate outward FDI of Brazilian firms.

This paper is organized as follows: Section 2 puts Brazil’s BIT policies in context and it introduces two claims made by scholars regarding the impact of these BIT policies on FDI. Theoretical (economic, political science, and legal) justifications of BITs are summarized in Section 3. Section 4 reviews the historical genealogy of BITs in Brazil. Section 5 provides an overview on the established empirical evidence of four types of effects of BITs on FDI, and it reflects on the Brazilian transition to the new type of BITs in the light of this evidence. Section 6 summarizes, and Section 7 sketches limitations and avenues for further research.

2. Brazil’s BIT Policies in Context

Brazil has signed several BITs between 1994 and 2020. These include 14 traditional BITs between 1994 and 1999, 13 CFIAs between 2015 and 2020, and 20 treaties with investment provisions (thereof 16 in force)—most of them related to MERCOSUR.4 In addition, Brazil has adopted 20 investment-related instruments.

Only three BITs, both from the second period (2015–2020), have been ratified and thus are actually in force.5 Monebhurrun (2017, p. 100) stresses that these agreements “symbolize Brazil’s true entry in the landscape of international investment law”. Carvalho (2018, p. 33) remarks concerning the treaties of the second period that “the signed agreements seem to be following the same path towards ratification as the two already ratified agreements”.

It is for sure that in recent years, other countries have taken different unilateral and multilateral routes to IIAs (see Behn et al. 2015; St. John 2018, p. 239;

Table 1. Country-level reactions to critical issues about IIAs on the national level.

Denouncing the system of BITs and of Investor–State dispute settlement (ISDS; e.g., South Africa)

Not renewing their BITs (India, see Kotyrlo and Kalachyhin 2023)

Following individual trajectories (e.g., Brazil and its treaty partners (back to State–State dispute settlement); China (new dispute settlement mechanism ISDS))

Withdrawing from the system (e.g., Venezuela in 2012, Bolivia since 2009, Ecuador from 2009 to 2021) and the denunciation of the Convention on the Settlement of Investment Disputes between States and Nationals of Other States, which constituted ICSID (Bas 2022)

Terminating BITs (e.g., India’s unilateral mass termination); gradually terminating BITs (e.g., Indonesia; Pakistan, which terminated 23 out of 48 BITs and will not ratify 16 signed BITs)

Considering moving or actually moved away from the Energy Charter Treaty (in 2023, the European Commission (EC) formally recommended a coordinated European Union (EU) withdrawal from the Energy Charter Treaty6; see United We Leave or Divided We Stay? Why it’s time for the EU to speak with one voice regarding the Energy Charter Treaty | International Institute for Sustainable Development (iisd.org accessed on 20 May 2024))

Expanding the role of exception clauses (e.g., Brazil)

USA phasing out its ISDS with CAN in the new NAFTA (USMCA—United States–Mexico–Canada Agreement)

Shifting to the mega-regional level: Integration of investment issues into trade agreements (e.g., the Regional Economic Comprehensive Partnership)

Source: Authors’ compilation.

below), and they are united in their discontent with the current BIT system. The list of states that sometimes radically change their position towards traditional BITs continues to grow (Soopramanien 2017, p. 609). Yet, the route that the Brazilian state has taken, that is back to a State–State dispute settlement (SSDS) and expanding exception clauses, certainly is of special economic interest.

Regarding the switch of Brazil to CFIAs, two fundamental claims have been made in the literature:

Claim #1: The switch to CFIAs and away from traditional BITs is justified as the non-ratification of the traditional BITs has not harmed FDI in Brazil.

This claim puts into question the purpose of traditional BITs. For example, Vidigal and Stevens (2018, p. 486) stress that “at least for a country with the characteristics of Brazil, it appears that IIAs are not determinant for attracting FDI”. Cavallo (2019, p. 69) mentions that “Brazil has consequently been seen as an example that BITs do not have any effects on FDI inflows, or that countries can do without them”.7

Claim #2: While CFIAs avoid some of the problems (i.e., direct and indirect costs) of traditional BITs, on balance they are less effective than traditional BITs would have been had they been ratified (e.g., Hegdekatte 2023; Sarmento 2021).

In what follows, we review the theoretical and empirical literature and investigate the empirical validity of these two claims. With regard to claim #1, empirical evidence in general as well as specific to Brazil suggests that Brazil has forgone FDI by not ratifying traditional BITs. The CFIA partner country mix and recent estimates of a rather low impact of BITs on FDI suggest that one should not be overly optimistic with regard to CFIAs as a promotor of FDI.

Concerning claim #2, while CFIAs include an alternative dispute resolution mechanism, which aims at a better compliance with a BIT’s rules, rather than the compensation of foreign investors, the lower stringency of the dispute resolution mechanism makes CFIAs less effective. Yet, this weakening effect included in CFIAs must be weighed against potentially positive effects on FDI from innovative clauses (such as environment, labor, human rights, and investor obligations), which are absent in many traditional BITs.

3. Theoretical Justifications of BITs: A Review

The overarching motive for the conclusion of BITs is to attract FDI. BITs can be separated into two types: BITs for FDI protection and BITs for FDI liberalization. Most BITs that entered into force in the 20th century were of the protection type, while the liberalization type has become the mainstream in the 21st century (Urata and Baek 2022). The Brazilian government has recently added a third type, which may be termed the facilitation type. While the protection type has a focus on the protection of property rights of foreign investors, the liberalization type emphasizes market access (admission and establishment; see Dolzer and Schreuer 2008, p. 79f.) as well as the protection of investors and investments in the pre-entry stage. The facilitation type aims at fostering the compliance with the treaty (i.e., on deterrence of a breach).

Common to all BITs is the inclusion of property rights protection clauses and procedural clauses. Property rights protection clauses may include, for example, the Most Favored Nation (MFN) clause, the Fair and Equal Treatment (FET) clause, or a Non-Discrimination (ND) clause. Procedural clauses typically include some form of dispute resolution mechanism, either an Investor–State dispute settlement (ISDS) or State–State dispute settlement (SSDS). From the viewpoint of foreign investors, the dispute resolution clauses are at the heart of BITs, as they include procedures as to how disputes between the investor and the host country government can be resolved outside the domestic legal system. While in international law, SSDS is standard, IIAs have increasingly shifted to ISDS.8 Various disciplines provide a range of theoretical justifications for BITs.

3.1. Economic Justifications

From an economic perspective, a credibility problem (due to information asymmetry) exists between the foreign investor and the developing host country government. Not all information about future government policy is publicly available and, therefore, the investor considers the possibility of opportunistic behavior of governments, once the investment is undertaken. This creates the so-called hold-up problem (political risk; in the context of investment treaties, Odoemena and Horita 2017; Bonnitcha et al. 2017; Horn and Norbäck 2019). The idea of reneging conforms with the obsolescing bargain view, where the ex ante power of the investor, who promises positive effects on development, shifts to the host country government, which may renege on the agreement ex post once the assets have been invested (“sunk”) in the host country.

In other words, “one possible role (for BITs, authors) is to serve as a commitment device for governments that lack the ability to make credible unilateral commitments to protect inward investment” (Horn and Norbäck 2019, p. 14). Cavallo (2019, p. 72) summarizes the economic view stating that “a host government’s credibility to abide by the terms of a contract increases when it enters into a BIT, which then stimulates greater FDI inflows”.

From a theoretical perspective, how do BITs provide a possibility to solve the hold-up problem?9 Broadly speaking, when they contain a clause, which makes ex post reneging on the agreement by the government irrelevant for investors, because they are compensated (i.e., investment neutrality). If the clause provides for full compensation of losses after expropriation, this reduces the political risk of a foreign investor and, hence, is an efficient solution to the hold-up problem (e.g., Guzman 1997). The above-mentioned ISDS clause is a clause in question, as it provides for a dispute settlement mechanism outside the domestic legal system of the host country. Indeed, this clause is considered “the most important aspect of the international protection of investments’” from the viewpoint of foreign investors (Schreuer 2013).10 Frenkel and Walter (2019, p. 1316) mention that “the literature argues that international dispute settlement clauses are a crucial component of BITs, as they provide a way to sanction deviating behavior, determine the credibility of legal promises and allow investors to enforce their rights independent of the local level of the rule of law”.

In addition to the commitment effect, some authors argue for a signaling effect of BITs—Cavallo (2019, p. 72): “(B)y signing BITs, host governments signal their underlying intention not to interfere with foreign investment”. When a country ratifies a BIT, “it sends a signal to investors from all foreign countries, including those not covered by investment treaties, that it has laws and policies in place that protect foreign investment.” (Bonnitcha et al. 2017, p. 158). And, from a practical economic perspective, Maggetti and Moraes (2018, p. 301) argue that “many countries signed up to BITs containing ISDS clauses—especially developing countries—as this was considered an important signal by cash-strapped economies to attract much-needed foreign investment”.

To act as a credible signal, it must be less costly for states with favorable investment climates “to become and remain parties to investment treaties than it is for states with bad investment climates. If that were not the case, all states would be equally likely to ratify investment treaties, and the existence of a treaty would not convey any meaningful information about the quality of the investment climate” (Bonnitcha et al. 2017, p. 158). For the credibility of the signal, thus, ISDS again is crucial. For states with inferior investment conditions, i.e., high political risk, the probability to face costs arising from ISDS claims is high, while for countries with good investment climates, the risk is miniscule.

Haftel (2010, p. 359) points out that only signing, but not ratifying, a BIT does not send a positive signal to investors: “since signing a BIT is not costly, foreign investment cannot rely on this action to distinguish between pro-FDI and anti-FDI countries. In this interpretation, signing a BIT does not add believable information to foreign investors. A signed BIT does not protect investors in situations in which the host government violates their legal rights. It is not surprising, then, that signed treaties fail to boost FDI”.

In addition, signing but not ratifying could be seen by foreign investors as evidence that the host government does not intend to honor the terms of the treaty and that the rights of foreign investors enshrined in the treaty are not fully guaranteed, or even that the intention of the host country is not to provide the rights of the signed treaty to foreign investors. Thus, while the commitment effect is relevant only to foreign investors covered by the BIT, a possible signaling effect is relevant to all foreign investors.

3.2. Political Science Justifications

From a political science perspective, the rationales of home countries of FDI to ratify BITs are the de-politization of disputes and the reduction in diplomatic interventions, which may jeopardize foreign policy goals (Jandhyala 2015). However, Gertz et al. (2018) find no evidence of the de-politization of dispute resolution through BITs.

Another political science-related justification of BITs is expressed by Ostřanský and Aznar (2023), who argue that “conventional accounts presume that these treaties improve national governance, leading to good governance and the rule of law for all”. Yet, critical accounts charge that investment treaties empower foreign investors and cause a regulatory chill.11 Indeed, the most compelling caveat to BITs from the government’s point of view concerns regulatory chill as a reaction to investment disputes and this pertains to all areas of economic activity (Berge and Berger 2019; Horn and Norbäck 2019). Horn and Norbäck (2019, p. 26) even argue that regulatory chill is “one of the most pervasive claims” in the debate about BITs. They, however, also mention that there is little systematic empirical evidence to support this claim.

Concerning the possibility of regulatory chill, ISDS is again of special importance. According to Bas (2022, p. 3), the ISDS mechanism “acts as a filter that constrains national policy space by allowing foreign investor to question measures taken in exercise of the right to regulate (the heart of the policy space)”. Due to ISDS, “investment courts operate as external control bodies of the legality of States’ actions, even regarding human rights, public health or environmental protection” (Bas 2022, p. 2).

3.3. Legal Justifications

From a legal point of view, the provision of legal standards for an investor-friendly environment aims at “safeguarding foreign investment against interference of the host state” (Schreuer 2013). To this end, a BIT defines “obligations of the host states to treat foreign investment according to specific standards” (Bonnitcha et al. 2017, p. 11).

It is important to note that international investment law “sets accepted standards for the unilateral conduct of the host state” (Dolzer and Schreuer 2008, p. 22). The provision of legal safeguards includes the stability of the legal conditions under which the foreign investor can operate, the quality of the local public administration, the transparency of the system of local regulations, and an effective system of dispute settlement (Schreuer 2013). The latter is backed by a theoretical argument, initially advanced by Allee and Peinhardt (2014, p. 60), who maintain that “if the host has a particularly strong need to convey its intent to respect FDI, it is more likely to sign treaties that contain elements to strengthen enforceability, including preconsent clauses and multiple, institutionalized enforcement options”. These legal safeguards comprise domestic laws and IIAs.

Furthermore, from a legal perspective, investment treaties respond to the bias and unreliability of domestic courts. Hence, concerning arbitration, it is claimed that investment arbitration advances fairness and the rule of law in the resolution of investment disputes (Van Harten 2010). In particular, legal scholars emphasize the increased efficiency of the process of investment disputes and the final execution of verdicts (no legal recourse) as well as transparency.

None of these legal justifications have, however, fully materialized. There are major concerns about the efficiency of the process, as the average length of decided cases is 3.73 years (Behn et al. 2020, p. 209). Also, the final execution of verdicts poses tremendous problems in some of the disputes, and Pohl et al. (2012, p. 36) note that “in only 35 out of 1600 treaties examined, the verdicts of the arbitration councils have to be published”.

Table 2. Summary of main motivations.

Approach	Motivation for IIAs is to increase FDI (inward and outward) via...
Economic	Commitment effect
	Signaling effect
Political Science	De-politization and reduction in diplomatic interventions
	Improvements in national governance
Legal	Provision of legal safeguards
	Increasing the efficiency of the process
	Final execution of verdicts
	Transparency of process and reasoning

Source: Authors’ compilation.

summarizes the main motivations for the conclusion of BITs. In most of these motivations, an effective ISDS mechanism plays a key role.

We now turn to a discussion of Brazil’s position towards BITs over time.

4. Brazilian Bilateral Investment Treaties: An Overview

This section reviews the transition from the early stages of BIT conclusion up to the present situation in Brazil.

4.1. The Early Stage—1994–1999

Concerning the traditional BITs, which Brazil did not sign before 1994,12 Escobar (1996, p. 91) mentions that “one of the most prominent features of these and of all recent investment treaties (of Brazil, authors) is their provision for the settlement of disputes between covered investors and the host state by means of binding international arbitration”.13 Regarding conflict resolution enshrined in these agreements, it is worth noting that the BITs of the early stages have shifted from SSDS to state–private entity dispute resolution (Maggetti and Moraes 2018).14

Between 1994 and 1999, Brazil signed 14 traditional BITs. Yet, “they were never approved by the country’s National Congress, which saw the investor–state arbitration regime as limiting states’ right to regulate15 and as granting extraordinary benefits to foreign investors, hence discriminating against domestic investors. For the same reasons, Brazil did not sign the 1965 Convention on the Settlement of Investment Disputes between States and Nationals of Other States (ICSID Convention).16 Even so, it continued to receive significant amounts of foreign direct investment (FDI), consolidating its position as one of the world’s top recipients of FDI and reinforcing the understanding that having BITs in force is not decisive for attracting investments” (Martins 2017, p. 1)17.

Despite that BITs were part of Brazil’s development strategy, and despite Brazil’s quest to attract FDI mentioned above, Welsh et al. (2014, p. 112) explain that “Brazil’s negotiations over BITs, however, must be understood as largely reactive. In most cases, the home states of international investors initiated the negotiation process for Brazilian BITs as they sought to assist their corporations with managing costs and risk”. This also explains why the early BITs negotiated by Brazil very much adhere to the model of developed countries, including ISDS provisions (e.g., Brazil–Germany BIT, Brazil–UK BIT).

At this point, it is important to understand that during the 1990s when Brazil’s BIT negotiations started, several economic and political factors were given that may have impacted positively on FDI and which help to explain the non-prioritization of BITs besides the direct criticism of being unconstitutional:

• Brazil’s executive and legislative authorities created alternative sources of investment protection (Cavallo 2019, p. 71), so the rejection of one type of investment protection was accompanied by another type of investment protection. “Major elements of investment protection identified as international standards […] were addressed by the Brazilian government during the 1990s and early 2000s” (Welsh et al. 2014, p. 123). This “met the most important needs of foreign investors while protecting the most important interests of the state, thus substantially eroding the perceived need for a BIT”. (ibid., p. 123)18.
• With regard to investment disputes, Brazil built a “local or national-level system of investment arbitration that differs from the ICSID model in terms of language, location, the requirements for consent, and subject matter jurisdiction” and thus, “even without BITs and a system of ICSID- style quasi-precedential arbitral decisions, Brazil provides the same level of investment protection” (Welsh et al. 2014, p. 129).
• The BIT policy was not a stand-alone policy, as Brazil continued to introduce a number of important policy measures conducive to the attraction of FDI.19
• Brazil joined MIGA in 1993, giving full access for foreign firms to political risk insurance.

As a result of the interplay between political and economic factors as well as legal developments, the traditional BITs signed in the early period (i.e., 1994–1999) were very much in line with the treaties that other Latin American countries concluded and indeed have been negotiated by countries in other parts of the world (see also Welsh et al. 2014, p. 109). This is an expression of the fact that the negotiating partners were powerful “northern” countries and also of the convergence of the BIT model during that period.

Eventually, in 2002, the executive branch withdrew the signed agreements from Congress after determining that “treaty provisions on international Investor-State Dispute Settlement (ISDS) were unconstitutional” (US Department of State 2023, p. 12). Titi (2016, p. 6) explains: “Another element that at least at some stage seemed to reinforce Brazilian discomfort with the ICSID Convention was the fact that the latter signified a step away from diplomatic protection. What in other states may have played out in favour of the investor-state dispute settlement mechanism, in Brazil seemed to constitute an argument for its rejection”. In addition, Monebhurrun (2017, p. 82) mentions the “lack of equality and reciprocity between the host State on one side and the investors and their home States on the other”.20 “Investor-state dispute settlement provisions […) were perceived not only as unconstitutional […) but also as contrary to Brazil’s wish to adopt public policies critical to the country’s development” (Titi 2016, p. 6)21.

Welsh et al. (2014) and Campello and Lemos (2015) provide a detailed account of the underlying political process behind the non-ratification. Out of the three main potential factors for the non-ratification of the BITs, none actually apply to Brazil according to Campello and Lemos (p. 1064f.): (i) the absence of presidential veto power; (ii) constraints on executive power; and (iii) content of particular treaties.22 Campello and Lemos (2015) argue that “the main factor to explain the quandary of non-ratification was an unresolved executive, which was never fully committed with the treaties in the first place, and became less and less so as their costs and benefits became clearer over time”. (p. 1066). One could say that this is a case of a learning process in the government on the basis of concerns raised by the opposition.23

In addition, international developments became relevant as Argentina experienced a large number of investment disputes related to emergency measures introduced during the Peso crisis (see Bellak and Leibrecht 2021). Perrone and César (2015) point out, “the 2001 Argentine crisis confirmed Brazil’s concerns about investor-state arbitration and national sovereignty”. (p. 1).

4.2. The Transition to New Agreements

Why has Brazil started new BIT negotiations after a long period of inactivity and why is the result of these negotiations so different to the BITs of the early stage? Already in 2014, Welsh et al. (p. 107) hypothesized that “(m)eanwhile, Brazil’s role in foreign investment has evolved as its own multinational corporations increasingly engage in foreign investment. Inevitably, these corporations seek to reduce the risk of their foreign investments. As a result, they may encourage Brazil’s executive and legislative branches to take a second look at BITs. As circumstances change, so may the definition of success”.

This movement is noted in the early 2000s, in which “a transition from the Brazilian position as solely a large investment recipient to also being an investor has led to the adoption of a new standard of investment protection for Brazil, the CFIA (…)” (Gabriel 2016, p. 156).

Maggetti and Moraes (2018, pp. 301, 309) refer to the transition process as a policy learning process of a late adoption of the traditional policy preference of Brazil, which can be referred back to the fact that it “has been proving the right one for a country like Brazil”. Moraes and Cavalcante (2021, p. 309) describe the genealogy of the new agreements as reflecting “the concerns voiced during the BIT debates of the late 1990s–early 2000s”.

As a result of the political debate, Brazil’s new model treaty of 2015 paved the way to today’s CFIAs,24 and the next subsection refers to the features of this new type of investment treaty.

4.3. The Current Stage—Post 2015

After a period of sixteen years (from 1999 to 2015), Brazil entered into new bilateral treaties, which were inter alia based on the new model BIT of 2015, the CFIA.25 Generally speaking, “the new Brazilian treaties strengthen state politics in FDI relations” (Perrone and César 2015, p. 2).

By the year 2020, already 13 CFIAs were signed and 2 of them are ratified. Carvalho (2018, p. iii) argues that three factors have been crucial in bringing the success in ratification of this latest wave of agreements: “(1) the reformulation of the investment agreements from the traditional template of BITs to the new template of Cooperation and Facilitation Investment Agreements (CFIAs); (2) the changing role of Brazil from an importer to both an importer and an exporter of capital; and (3) the extensive political organization, both inside and outside the government, to support the ratification of the agreements”. Interestingly, the mix of CFIA partner countries neither represents major current trading partners nor major current FDI partners of Brazil (see Perrone and César (2015) for possible explanations of this fact).

The new CFIAs differ in form, substance, and procedure (Roberts 2018) from the traditional BITs. Regarding “form”, Brazil still sticks to the bilateral type of agreements, rather than regional or multilateral types (see

Table 3. Reform matrix, Brazil.

	Incremental	Systemic	Paradigmatic
Form	Brazil
Substance		Brazil
Procedure			Brazil

Source: Authors’ compilation, based on Roberts (2018).

). Regarding “substance”, Brazil’s BIT reform entails a systemic change, where “reforms work within the current system but require larger-scale, often-structural reforms” (Roberts 2018, p. 191). With regard to “procedures”, Brazil even introduces a new paradigm for conflict resolution.

Taken together, these reform elements justify calling the CFIAs a “new type of BIT”.

Table 4. Omitted and included clauses, relative to traditional BITs.

Contents	Omitted Clauses in CFIAs	Included Clauses in CFIAs
Rights protection	FET	The transparency clause may be “construed as another means of bringing the fair and equitable treatment standards, at least some of its elements, into the debate”. (Monebhurrun 2017, p. 94)
	MFN	MFN (with qualifications)
	Full protection and security	Transfers (temporary exceptions: balance of payments crisis)
	Tax measures	Provisions on Investment and Environment, Labor Affairs and Health
	Prudential measures	NT (with qualification)
	Protecting the right to regulate Safeguarding the ‘right to regulate’ by avoiding open-ended language that could be interpreted in ways that curtail the State’s policy space; carve-outs from the coverage of their obligations to investors. Exceptions are (i) public policy exceptions and (ii) security exception. Based on Moraes and Cavalcante (2021, p. 12)	Only direct expropriation is covered (indirect expropriation is explicitly not covered)
Procedure	ISDS	National Focal Points or Ombudsman
		Joint Committees (for the administration of the treaty)
	-	Dispute Prevention Procedure
Cooperation facilitation	-	Voluntary corporate social responsibility
	-	Exchange of information between parties; “multiple stakeholder dialogue” (Monebhurrun 2017, p. 86)
	-	Investor obligations
	-	Compliance with laws
	-	Environmental clauses (see Martini 2017): Model BITs that do contain at least one reference to the protection of the environment or sustainable development—Brazil (2015). FN 179: The similarities between Brazil’s model BIT and recently negotiated treaties such as the Brazil–Chile BIT (2015) Only preamble, no legal power
		Civil society participation
		Agenda for Further Investment Cooperation and Facilitation

Source: Authors’ compilation. Note: If clauses appear in both columns, it means that some CFIAs contain the clause, and some do not. Mainly based on the Brazil–India agreement; the items in italics are considered innovative clauses.

sketches the main differences between traditional BITs and CFIAs (also see Vidigal and Stevens 2018, p. 487).26

In summary, the most remarkable and distinguishing feature of CFIAs in the context of this paper is the shift from a state–private entity dispute settlement (i.e., ISDS) back to SSDS.27 For example, the Investment Cooperation and Facilitation Treaty Between The Federative Republic of Brazil And The Republic of India, para. 19.2, states: “The purpose of the arbitration is to decide on interpretation of this Treaty or the observance by a Party of the terms of this Treaty. For greater certainty, the Arbitral Tribunal shall not award compensation”.

How do we evaluate this shift to SSDS from a theoretical viewpoint? In order to solve the above-mentioned hold-up problem, according to Bonnitcha et al. (2017, p. 136), a BIT clause must contain the following: “First, it needs to impose an obligation on the host state to pay compensation if it expropriates an investment or reneges on an agreement relating to the investment. Second, this obligation needs to be enforceable in practice. Third, the compensation paid if the host state breaches the obligation needs to be calculated by reference to the loss suffered by the investor. (If the third condition is not satisfied, an investor might successfully demand compensation for expropriation of its investment, but still be left worse off than if it had not made the investment)”.

Dispute clauses included in CFIAs fulfill none of the three conditions above. From a theoretical point of view, as the CFIAs include SSDS provisions instead of ISDS, due to their different nature, these are not able to solve the hold-up problem in the same way, since currently in the Brazilian CFIAs, the compensation of investors is explicitly excluded (or based on SSDS at best). But, if the traditional dispute settlement were included in CFIAs as well, ISDS could have a chilling effect (Tietje 2014), which works against the FDI promotion effect. Hence, from a host country perspective, CFIAs without ISDS may be superior, as the risk of abending the introduction of welfare enhancing policy measures in the public interest would be lower. Moreover, under CFIAs, the Brazilian government would not necessarily have to pay compensation, of course depending on the outcome of the precise SSDS process in each investment dispute.

The move of Brazil from traditional BITs to CFIAs is quite surprising as still, in 2023, Yarygina and Krylova (2023) emphasize that investing in Brazil remains risky and hence the protection of property rights should be considered important in the view of foreign investors. Indeed, The World Economic Forum28 ranks “failure of national governance” on top of the business risks for Latin America in its Executive Opinion Surveys in several years.

In the debate surrounding ISDS vs. SSDS, one should also not forget that the shift from ISDS to State–State conflict resolution may cause diplomatic problems between states (especially when larger firms are involved) and hence, this will also not be without costs.

In short, compared to traditional BITs, the new CFIAs include some traditional and some rare, new, and innovative clauses and exclude some traditional clauses, notably ISDS and FET. The omission of an ISDS clause is interesting as while the basic motivation of the Brazilian government to sign CFIAs instead of traditional BITs is again to stimulate increasingly outward FDI, the path of Brazil is contrary to other large countries like China, which transitioned from a non-ISDS (or limited ISDS) to a strong ISDS status in order to incentivize outward FDI (see Leibrecht and Bellak 2023).

5. Empirical Effects of BITs on FDI: A Review

This section scrutinizes the effects of BITs on FDI of other countries on three different levels: the conclusion of BITs (4.1), the inclusion of ISDS (4.2), and the effects of ISDS claims on FDI (4.3).

5.1. The Effects of the Conclusion of BITs on FDI

In a widely quoted study, Brada et al. (2021) use 2107 estimates of the effect of IIAs on FDI drawn from 74 empirical studies. Adjusting for publication bias, their meta-analysis “finds robust evidence that the effect of international investment agreements is so small as to be considered zero”. (p. 1). They state that the literature covered by their meta-analysis reports “trivial effects of IIAs on FDI”. (p. 24). A similar conclusion is reached by Reiter and Bellak (2020) upon 721 semi-elasticities drawn from 40 studies. Their findings “point to the existence of a positive publication bias in subgroups of the studies, suggesting genuine effects of BITs on FDI below 1%”. (p. 1).

Falvey and Foster-McGregor (2018, p. 17), a study not covered in the meta-analyses, on the basis of a difference-in-difference analysis, using the sample of treated and matched control observations obtained using the matching analysis, find that “while the coefficients on the BIT variable are positive and economically large in the matched sample, they are also insignificant”. Estimating the intensive and extensive margins for country pairs, their results show that “forming a BIT between country-pairs where FDI was already growing has at best an insignificant effect and at worst a negative impact upon FDI flows. In contrast, we find large, positive and significant coefficients on the BIT dummy in the case of BIT events along the negative intensive margin—i.e., in country-pairs that formed a BIT after a period of disinvestment” (Falvey and Foster-McGregor 2018, p. 17). Although the effect is in their words “outlandishly” large, they contend that “it does suggest at least that large reversals in FDI outflows can be achieved if investors perceive a positive change in the environment”. For the extensive margin, their results imply “an increase in FDI flows along the extensive margin of between 71 and 113%” (Falvey and Foster-McGregor 2018, p. 18).

For the specific case of Brazil, Cavallo (2019) provides an interesting counterfactual analysis of the impact of BITs on FDI in Brazil. Cavallo (2019, p. 71) hypothesizes that “Brazil might have been the main destination of FDI in Latin America for the past couple of decades despite not enacting a single BIT, yet this does not mean it was achieving its full potential as FDI magnet”. Cavallo (2019) applies the Synthetic Control (SC) approach of Abadie et al. (2010) and finds that if Brazil had ratified BITs in the early 1990s, it would have received even more FDI. He concludes that Brazil “missed out on a significant amount of FDI inflows due to its position towards BITs”. (p. 96).

We cannot assess here how serious the hold-up problem is for the case of Brazil, as hold-up issues vary and evolve over time, e.g., with political changes (a change in government or the presidency after elections), economic crises, etc. In any case, the results put forward in the meta-analyses and the earlier studies should not be viewed as evidence of whether investment treaties are effective in resolving the hold-up problem (Bonnitcha et al. 2017). Since BITs have been evaluated in their entirety in the underlying studies, the resulting effect on FDI must not be referred to the effect of a single clause, like ISDS, on FDI. This is also the case for the Brazilian study by Cavallo (2019) as his analysis cannot trace the negative effect to the existence or absence of an ISDS clause in a BIT as the units of his analysis are Brazil’s BITs in their entirety. Studies investigating the specific effects of ISDS on FDI are summarized next.

5.2. The Effects of the ISDS Clause on FDI

How does the inclusion of an ISDS clause in an investment treaty affect FDI empirically? The difference an ISDS clause makes in the effect of FDI should be interpreted as a rough indication of what the exclusion of ISDS from CFIAs may imply for FDI inflows, as the effects of the inclusion and the exclusion of such a clause need not be symmetric. The results range from a minor role (e.g., Berger et al. 2013) to a key role (Desbordes 2016); the latter author remarks: “In the case of BITs, the presence of an investor-state dispute mechanism is the only provision which appears to matter”. (p. 23).

A particularly interesting study in this respect is by Dixon and Haslam (2016), who point out that early studies about BITs’ effects on FDI suffer from several issues. Many of those can be summarized as a lack of differentiation of the BITs, among them the use of “imprecise proxies for treaty strength”. (p. 1081). Therefore, Dixon and Haslam (2016) propose to assess the effect of ISDS on FDI by differentiating treaty clauses and by including other treaties with investment provisions. What is notable about their study in our context is their focus on Latin American countries.

Dixon and Haslam (2016) use a variable that measures the degree of treaty protection offered to foreign investors and which, according to the authors, enables them to distinguish the signaling effect and the commitment effect of BITs empirically. They measure the overall strength of the BIT, rather than specifically of ISDS. Dixon and Haslam (2016) find that “ratification offers no additional impact on FDI flows over and above signing” for N-S (north–south, authors) FDI (p. 1095) and they “find no evidence of an IIA effect in the north–south sample, reflecting either a signaling or credible commitment, on FDI flows”. (p. 1095).

Concerning south–south (S-S) FDI, Dixon and Haslam (2016) find that only strong treaties do impact on FDI, which is not encouraging given that Brazil has signed CFIAs primarily with S-S countries. Indeed, CFIAs fall into the category of “weak” treaties, where Dixon and Haslam (2016) established a statistically insignificant effect on FDI.

The results of Frenkel and Walter (2019) again indicate that “stronger international dispute settlement provisions in BITs are indeed associated with more FDI activity”. Using various FDI measures and an index of “BIT strength” by differentiating different degrees of dispute settlement clauses, they find in a bilateral setting that “for developing host countries, on average an increase of the BIT index of one point leads to a 5.1% increase of FDI inflow from the partner country”. (ibid., p. 1335). This includes a two percentage point lesser increase when all countries are included.

Like Dixon and Haslam (2016) as well as Frenkel and Walter (2019), Ahmad et al. (2022) also distinguish between different strengths of BITs, but they focus on ISDS29 and, hence, the paper is especially useful for comparing traditional BITs and CFIAs. Ahmad et al. (2022) apply an augmented gravity model and “test the impact of several policy features that limit the scope of ISDS in order to discern whether restrictions on ISDS result in reduced FDI” (p. 1). For “strong” ISDS, they find a positive and large (22%) increase in FDI. More importantly, in light of the discussion on CFIAs, they find a reduction in FDI of 40% for agreements without ISDS. This finding is surprising not only because of the size of the effect, but also because BITs also impact through other channels on FDI.

Evidence for the importance of ISDS is also presented by three country-level (India and China) case studies. India terminated its BITs mainly due to concerns over the increasing number of ISDS claims it faces (Hartmann and Spruk 2023). Kotyrlo and Kalachyhin (2023) evaluate the FDI effect of the termination of BITs by India. They find that the termination leads to a significant decrease in FDI inflows from the partner country. Hartmann and Spruk (2023) unveil a significant reduction in FDI inflows to India by more than 30 percent when compared to countries without BIT terminations. Leibrecht and Bellak (2023) investigate a major BIT reform in China and note that the substitution of first- with third-generation BITs, whereby first-generation BITs include no or only limited ISDS and third-generation BITs include strong ISDS clauses, exerts a positive impact on the quantity of FDI received by China.30 The estimated coefficients of the variable indicating the ratification of a third-generation BIT suggest a “range from USD (constant 2010) 185 to USD 101 per capita. A conservative interpretation is that the impact of the BIT reform on inward FDI in China is moderate”. (ibid., p. 16). Similar arguments are provided by Du (2023) and Li and Bian (2020). Leibrecht and Bellak (2023) conclude that “(g)iven the economic importance of China, its political stability and regime durability, long-term orientation in policy as well as its informal channels of dispute resolution, one may ask why formal institutions like the more investor-friendly BITs have an impact on investors from treaty-partner states at all. The findings of this study are consistent with the view that the formality (enforceability) of investor-friendly BITs matters for FDI. Informal negotiations with public officials depend on their willingness to hear a claim. BITs of the third generation, in contrast, through ISDS, guarantee that public officials hear investors and their claims. Our finding of a positive impact of more investor-friendly BITs on FDI points towards the relevance of the enforceability of property rights for investments”. (p. 16). From this conclusion by Leibrecht and Bellak (2023), one might infer that the ratification of investor-friendly BITs (i.e., including strong ISDS) by Brazil, which is relatively less politically stable and less long-term-oriented, likely would have had a positive impact on FDI in Brazil.

Overall, the empirical studies are consistent with the view that ISDS does positively affect FDI. However, one must weigh these results with the possible negative impact of resulting disputes (ex post costs) to which we turn now.

5.3. The Effects of Investment Disputes on FDI

Allee and Peinhardt (2011) find “that BITs do increase FDI into countries that sign them, but only if those countries are not subsequently challenged before ICSID. On the other hand, governments suffer notable losses of FDI when they are taken before ICSID and suffer even greater losses when they lose an ICSID dispute”. Kerner (2009, p. 77) in addition mentions the “possibility of diminished credit rating” and “the (as yet never exercised) possibility of being referred to the International Court of Justice for non-compliance with an ICSID ruling”. (also see Rajput 2022).

In contrast, Minhas and Remmer (2018, p. 23) find no evidence of any effects of claims on FDI, but they find reputational effects for the post-2006 period. Their results put into question the “logic underlying the credible commitment story espoused in the political economy literature, which assumes that states incur statistically and substantively significant costs from allegations that they have violated their international agreements”. (p. 23). This result is contradicted by the findings of Aisbett et al. (2018, p. 119), who focus explicitly on developing countries and conclude that BITs “stimulate bilateral FDI flows from partner countries—but only so long as the developing host country has not had a claim brought against it to arbitration”. It is interesting to note that this effect increases with the number of claims in an average country.

Du and Zhang (2024) put the focus on the impact of ISDS claims on cross-border Mergers and Acquisitions (M&A). They find evidence that arbitration cases impact M&A negatively, especially in institutionally weak countries. Du and Zhang (2024) also unveil that it is the claims that are lost by governments, which lead to this reduction in M&A activity. Claims won by the government, in contrast, lead to “an acquittal effect that encourages the subsequent capital inflow to the respondent state”. (p. 1).

Again, however, these findings are disputed in a recent paper by Kerner and Pelc (2021), who claim that investors in recent periods are less likely to update their expectations on the basis of investment disputes and consequently do not reduce investments.

Taken together, the empirical literature is consistent with the view that ISDS claims have a negative impact on FDI in developing countries, but the effect may have diminished over time.

6. Conclusions

We study two claims that have been made with respect to Brazilian BITs, namely

Claim #1: The non-ratification of the traditional BITs has not harmed FDI into Brazil, which puts into question the purpose of BITs.

Claim #2: While newly introduced BITs, the CFIAs, avoid some of the problems (direct and indirect costs) of traditional BITs, on balance they are less effective than traditional BITs would have been.

This paper has reviewed some of the concerns that have been raised with regard to Brazil’s unique trajectory toward CFIAs. Hitherto, the literature on Brazil’s CFIAs mainly focused on legal issues and thus has largely neglected economic and political science issues.

On the basis of available empirical evidence of various types of economic effects of BITs, the following conclusions emerge:

Concerning claim #1, empirical evidence in general as well as specific to Brazil suggests that Brazil has forgone some FDI in the past. Cavallo’s (2019) analysis of Brazil shows that the inflow of FDI would have been substantially larger had Brazil ratified traditional BITs. No single factor has been found that would suggest that Brazil’s BITs would have a structurally different effect on inward FDI than in other countries. Generally, empirical evidence about the relation between BITs and FDI is mixed. Studies on the mere presence of a BIT show a modest positive impact on FDI, while studies analyzing the effect of the presence and strength of ISDS find larger impacts on FDI. Yet, the impact of this effect may be mitigated if foreign investors are prepared to take up political risk insurance instead, in order to be present in this important market.

Regarding claim #2, so far, no empirical evidence has been produced that CFIAs are less effective, but theoretical legal and economic arguments point in this direction. Theoretically, the incentive effect for outward FDI, which Brazil intended to support with the transition to CFIAs, is clearly reduced by excluding ISDS, but this must be weighed against the expected benefits of the new cooperation clauses. The fact that the mix of CFIA partner countries neither represents major trading partners nor major FDI partners of Brazil leads to the expectation that the impact of these agreements on inward and outward FDI will remain limited. In conclusion, our analysis rejects claim #1 but supports claim #2.

Choosing CFIAs over BITs must be seen as a deliberate and conscious decision by the Brazilian government. While the primary motivation of this paper has been the examination of claims made with respect to the new CFIAs, we would also like to point to several tentative policy conclusions, which follow from our findings regarding the validity of the two claims introduced.

First, international investors seem to highly value the enforceability of property rights at international private tribunals as they are enshrined in traditional BITs. Hence, one needs to separate the issue of rights granted to foreign investors from the issue of the enforcement of these rights. As has been shown in this paper, CFIAs are distinct from traditional BITs in both aspects, since property rights, which have been key in investment disputes (e.g., FET), and a key enforcement mechanism (ISDS) have been omitted at the same time. Hence, it is undisputed that governments send a negative signal to foreign investors; whether it will be as detrimental as an exit from the system or a termination of treaties, only time will tell.

Second, with respect to the possibility that CFIAs are a role model for a new generation of BITs, the answer is negative in the sense that CFIAs likely will not fulfill the primary goal of BITs, namely the attraction of FDI. Governments thus have to resort to other means of FDI attraction like the provision of infrastructure, stepping up investment promotion activities abroad, etc. On the basis of the inclusion of innovative clauses in CFIAs, there is hope that governments will not engage in a race to the bottom on a broad scale.

Third, the sole reliance on a State–State dispute settlement as an enforcement mechanism suggests that governments are aware of the potential costs/externalities of a State–State dispute settlement. As long as FDI is a major component of a country’s development strategy, the enforceability of property rights needs to be guaranteed. Of course, this could be achieved by plenty of other means rather than BITs like impartial and efficient domestic courts, setting up an effective system of private political risk insurance (which has been rather limited in the past), or extending the public system of political risk insurance (e.g., MIGA). Additionally, the development of a new (multilateral) system of investment protection could be considered, but there are no indications that this is actively contemplated by Brazil at present.

As a consequence, countries relying on CFIAs will experience the locational competition much stronger than countries relying on traditional IIAs. Hence, the need arises to proactively undertake economic reforms stemming from the transition to or the introduction of CFIAs.

7. Limitations and Avenues for Future Research

We conclude that by avoiding ISDS (and FET), Brazil likely will forego FDI. However, in this review, we only used the literature that is available in the English language, and we left out the related literature in the Portuguese language. We therefore may have missed studies that paint a different picture regarding the effectiveness of Brazil’s FDI policies for FDI. What is more, with this review paper, we do not provide our own empirical evaluation of CFIAs’ effectiveness. Future analyses, therefore, could be based on Cavallo (2019) and use the Synthetic Control approach of Abadie et al. (2010) to investigate whether the ratification of CFIAs by Brazil with partner countries (the treatment pairs) spur Brazilian inward and outward FDI from and to the partner countries, respectively. Certainly, whether the opportunity costs of having CFIAs rather than traditional BITs in force are of an economically meaningful size needs additional investigation.