PTAs and their benefits

To address our central question, namely, the extent to which LAC countries benefit from participation in multiple and sometimes overlapping trade agreements, we draw on the economics and international political economy literatures on the expected benefits of PTAs. PTAs liberalize trade among members, providing them with preferential access to their respective markets while also discriminating against nonmembers. In evaluating their welfare implications, economists traditionally focused on the balance between their “trade creation” and their “trade diversion” effects (Reference VinerViner 1950). Yet this early analysis has been adapted to recognize the various dynamic effects that PTAs may have in a globalized economy.

Economists, for example, have also considered the various mechanisms through which preferential trade liberalization could work to increase FDI inflows, thus enhancing attractiveness of PTAs to developing countries. First, by creating an extended common market, PTAs may attract firms from non–member countries that want to take advantage of export opportunities (Reference MedvedevMedvedev 2012).Footnote ⁷ Second, by reducing the cost of trading and thus facilitating the exchange of intermediate and finished products among parent firms and their foreign subsidiaries, PTAs are expected to promote vertical FDI among members (Reference EthierEthier 2001; Reference Levy Yeyati, Stein and DaudeLevy Yeyati, Stein, and Daude 2003).Footnote ⁸ Along these lines, Vallejo and Aguilar (Reference Vallejo and Aguilar2002) found that the North American Free Trade Agreement (NAFTA), the Andean Community, and the Central American Common Market (CACM) contributed to an increase in intraregional FDI inflows.

Finally, economists and political scientists have emphasized the political economy mechanisms through which PTAs may work to attract foreign investment. In particular, PTAs may act as “commitment mechanisms” that signal policy transparency and predictability and thus contribute to mitigating the so-called time-inconsistency problem inherent in foreign investment (Reference Büthe and MilnerBüthe and Milner 2008, Reference Büthe and Milner2014). Indeed, once an investment is made, a rational government faces incentives to breach the terms of the agreement, expropriate assets, or introduce regulatory changes that make the investment less profitable for the firm. Firms will anticipate this type of opportunistic behavior and will be discouraged from investing in certain countries unless they can be given some guarantees.Footnote ⁹ An extensive body of literature has looked at the role of bilateral investment treaties (BITs) in helping states solve this credibility problem. BITs specify several guarantees to foreign investors, most notably, the possibility of seeking third-party arbitration in international tribunals, while signaling commitment to market friendly policies. Yet, membership in other types of international organizations and agreements can also work to signal and strengthen governments’ policy credibility (Reference Dreher, Mikosch and VoigtDreher, Mikosch, and Voigt 2015). Thus, Büthe and Milner (Reference Büthe and Milner2008) argue that by joining PTAs, and other trade institutions like the General Agreement on Trade and Tariffs and the World Trade Organization (WTO), developing countries show their commitment not just to the reduction of tariffs and other barriers to trade but more generally to liberal economic policies.

This commitment enhancing effect is reinforced in the case of recent PTAs, which contain provisions on domestic regulatory issues. Many of these deep agreements include specific investment provisions, such as for example, commitment to the removal of barriers to international capital flows and relaxation of restrictions on foreign ownership and participation, which are expected to result in higher FDI inflows. Similarly, provisions on intellectual property, competition policy, and standards harmonization may also attract more investment by improving countries’ investment climate (Reference MedvedevMedvedev 2012). Along these lines, several studies have found a positive effect on FDI inflows not just of the presence and the number of PTAs but also of their depth (e.g., Reference Büthe and MilnerBüthe and Milner 2008; Reference Orefice and RochaOrefice and Rocha 2014; Reference Osnago, Rocha and RutaOsnago, Rocha, and Ruta 2015).

Yet some studies have emphasized the potentially negative consequences of the coexistence of multiple, overlapping trade agreements. The so-called spaghetti-bowl effect refers to the complications and increased costs resulting from diverging rules of origin in overlapping PTAs. The trade agreement signed between the United States and the countries in Central America (DR-CAFTA), according to Granados and Cornejo (Reference Granados and Cornejo2006), illustrates these difficulties. This agreement brought together three preexisting agreements, namely, the Central American Common Market, the agreement between the five CACM countries and the Dominican Republic, and the free trade agreement between the US and these six countries. Despite similarities among the three arrangements, they differed in important ways, such as their rules of origin regimes. Once DR-CAFTA was established, the overlaps between the new multilateral regime of origin and the preexisting ones introduced fragmentation, increasing transaction costs for private and public actors (Reference Granados and CornejoGranados and Cornejo 2006). According to Bown et al. (Reference Bown, Lederman, Pienknagura and Robertson2017, 137), in fact, “the increasing number of FTAs, and the complex network of rules of origin that arises from that increase, can accentuate the uncertainty faced by firms seeking to invest in foreign countries and can deter foreign direct investment into a country.” In line with these studies, a report by the Dominican Ministry of Industry and Trade assessing the impact of DR-CAFTA claimed that the agreement has had only a “moderate” effect of FDI inflows because “its implementation has been challenging for the Dominican industrial sector” (Reference Villanueva and PolancoVillanueva and Polanco 2010, 6).

The intervening role of state capacity

PTAs have several potential economic effects for the countries that join them. Not only do they promote trade among members; they may also attract FDI. Yet trade agreements, and especially deeper ones, also pose important challenges for governments, particularly in developing countries. While the literature reviewed recognizes the problems associated with multiple, overlapping agreements, little attention has been paid to the importance of state capacity and domestic institutions in conditioning the impact of these agreements. On the one hand, state capacity conditions the ability of governments to negotiate and conclude trade agreements, thus influencing the number and nature of agreements that a country will belong to. But at the same time, extensive administrative, regulatory, and technical capacities are required not only for the negotiation but also for the implementation of trade agreements (Reference Chauffour and KleimannChauffour and Kleimann 2012). As Walls, Smith, and Drahos (Reference Walls, Smith and Drahos2015, 2) put it, “The capacity needed is expensive, skill-intensive, and requires considerable infrastructure, which smaller and poorer states especially struggle to find.”Footnote ¹⁰

Institutional capacity demands are heightened in the context of overlapping trade agreements. First, the simultaneous negotiation of multiple agreements requires extensive technical capacity and coordination across the different venues. For countries in LAC, the proliferation of PTAs with different countries within and beyond the Western Hemisphere increased the number of summits and meetings that public officials must attend, thus putting additional strain on national budgets (Reference Malamud and GardiniMalamud and Giardini 2012; Reference Mace, Thérien, Tussie and DabèneMace, Thérien, and Tussie 2015). The broader scope of many of these arrangements also called for engagement of multiple ministries and governmental agencies and their coordination among themselves and with other economic actors, including business groups and unions (Reference Heidrich, Tussie and StegerHeidrich and Tussie 2009). In addition, as countries become members of several agreements that differ in their sectoral coverage, liberalization schedules, and rules of origins, the challenges associated with their implementation increase not only for private actors but also for public officials in charge of enforcing them, as the example of DR-CAFTA discussed above illustrates (Reference Estevadeordal and SuominenEstevadeordal and Suominen 2009).

Indeed, for LAC countries, the administrative and technical costs of compliance and implementation grew dramatically with the signing of more comprehensive agreements with the US and other developed countries (Reference GonzálezGonzález 2009). These North-South agreements, which tend to reflect the preferences of industrialized partners, present complex implementation challenges, placing heavy burdens on developing countries (Deere 2005). While shallow agreements require relatively simple trade policy changes—for example, the reduction or elimination of tariffs—the more recent agreements involve substantive domestic regulatory and institutional reforms. Chauffour, González, and Maur (Reference Chauffour, González and Maur2011) identify several implementation challenges of deep integration agreements, all of which underscore the importance of governments’ institutional and regulatory capacities. First, deep PTAs may call for institutional changes, including the reorganization and/or creation of new agencies. In 2008, for example, a Ministry of Environment was created in Peru to meet the requirements of the US-Peru PTA (Reference Bastiaens and PostnikovBastiaens and Postnikov 2017). Second, some provisions, such as those concerning intellectual property rights and the environment, require the drafting of new domestic laws and regulations. In Central America, for example, Nicaragua and Costa Rica both amended domestic intellectual property laws to comply with DR-CAFTA (Reference Biadgleng and MaurBiadgleng and Maur 2011). Finally, these institutional and regulatory reforms entail important administrative and procedural changes, including training of staff, interagency coordination, and establishment of monitoring mechanisms.

According to Chauffour and Kleimann (Reference Chauffour and Kleimann2012, 49), the severe institutional and regulatory capacity constraints of developing countries constitute “the single most important obstacle to PTA implementation.” In line with this, we argue that by constraining the implementation of PTAs, poor regulatory and institutional capacities of states, undermine the ability of LAC countries to gain from deep trade agreements. Our analysis of the benefits of deep integration for Latin America and the Caribbean seeks to shed light on the intervening role of state capacity, understood as the ability of institutions to effectively implement a basic set of core functions and effectively enforce laws (Reference Luna and SoiferLuna and Soifer 2015).

An extensive literature in international relations has linked state capacity with implementation and effectiveness of international agreements (Reference Chayes and ChayesChayes and Chayes 1993; Reference GuzmanGuzman 2002; Reference ColeCole 2016). In contrast, the scholarship on PTAs has paid limited attention to the ways in which the administrative and institutional capacities of the state may affect a country’s ability to implement and hence benefit from these agreements. An exception is Gray (Reference Gray2014), who argues that the “implementation gap” that characterizes many regional trade organizations in the developing world is explained by problems of state capacity.Footnote ¹¹ Gray (Reference Gray2014) concentrates on the ways in which weak domestic legal environment and inadequate infrastructure hinder compliance and implementation and result in lower levels of trade interdependence. Relying on expert survey data, her work shows a strong correlation between the quality of domestic infrastructure and rule of law and the ability of regional arrangements to promote trade.

Like Gray (Reference Gray2014, 60), we believe that without sufficient domestic capacity, “the promise of international agreements may not always be fulfilled.” Specifically, we argue that by making compliance and implementation harder, weak state capacity also conditions the extent to which developing countries can reap the benefits of preferential trade liberalization. While Gray focuses on the effectiveness of trade agreements in promoting trade, we consider another expected their role in attracting investment by multinational firms.

There are two mechanisms through which state capacity may influence the potential of PTAs in promoting trade and FDI among member countries. First, and in line with Gray, we view state capacity undermining PTA effectiveness through its impact on implementation. In other words, by constraining compliance and implementation, weak state capacity and government effectiveness erodes the expected commercial benefits of PTAs. The inability of governments to stick to liberalization commitments and to implement harmonized rules and regulations will neutralize the ability of PTAs to promote intrapartnership trade. But in addition, weak compliance and implementation will increase other non-trade-related costs of transactions for firms operating in global value chains, creating incentives to shift production and investment to other markets. A second alternative mechanism is particularly relevant when considering the ability of PTAs to attract FDI. Inadequate rule of law and regulatory quality undermine the signaling and commitment enhancement effects of PTAs stressed by Büthe and Milner (Reference Büthe and Milner2008) and other international political economy scholars. As Gray (Reference Gray2014, 61) notes, “the implementation of international agreements requires that contracts of all types be enforced on the ground. Upholding the rule of law is another prerequisite for maintaining credible commitments.”

Along these lines, we hypothesize that a country’s state capacity affects its ability to benefit from trade agreements. We thus expect state capacity to significantly condition the effects of both the number and the depth of trade agreements that countries sign on FDI inflows:

Data and variables

To test these hypotheses, we compiled a data set on patterns of FDI and PTA membership for a sample of twenty-nine countries in LAC for the period 2001–2015. During this period, all countries in our sample have participated in at least one preferential trade agreement that has been notified to the WTO. The unit of analysis is the country-year and we use a fixed effect estimator, which allows us to control for any country time-invariant unobservable that may affect FDI and be correlated with agreement adherence and state capacity. We also include year effects to control for common shocks affecting all countries in the sample, such as global liquidity, or across-the-board liberalization trends. We estimate the following model:

$\mathrm{FDI} \mathrm{Inflow}{s}_{\mathrm{it}}={\alpha }_i+{\gamma }_t+{\beta }_1 \mathrm{PTA}{s}_{\mathrm{it}}+{\beta }_2\mathrm{State}\mathrm{Capacit}{y}_{\mathrm{it}}+{\beta }_3\mathrm{PTA}{s}_{\mathrm{it}}*\mathrm{State}\mathrm{Capacit}{y}_{\mathrm{it}}+{\beta }_4\mathrm{Depth}+{\beta }_5{X}_{\mathrm{it}}+{\epsilon }_{\mathrm{it}}$

The dependent variable (FDI_it) is the net FDI inflows into a given country in a given year, measured in millions of current US dollars. Data for this variable is from the UNCTAD database on FDI flows. There are three main independent variables of interest. The first two capture the effects of the number and depth of trade agreements. First, PTAs is the cumulative number of preferential trade agreements (in force) to which a country belongs at the end of a given year.Footnote ¹² In constructing this variable, we use data from the WTO’s regional trade agreement database. We include not only full-scope free trade agreements (FTAs) such as NAFTA but also partial-scope arrangements, like the Latin American Integration Association (ALADI). In addition, we include agreements signed with countries in other regions. The value of PTAs varies significantly across countries and years in our sample. In 2015, for example, the cumulative number of PTAs ranged from 1 for Aruba to 23 for Chile. In addition, the cumulative number of PTAs in force has grown rapidly in recent years, as shown by Figure 1.

Second, to measure the depth and scope of trade agreements, we use an index constructed by Hofmann, Osnago, and Ruta (Reference Hofmann, Osnago and Ruta2018) that aggregates the number of provisions in PTAs signed by each country. Drawing on Horn, Mavroidis, and Sapir (Reference Horn, Mavroidis and Sapir2010), this index identifies a set of policy areas covered in trade agreements, which can be classified according to whether they fall under the current mandate of the WTO (“WTO-plus”) or whether they are outside its reach (“WTO-extra”). Of the fifty-two types of provisions identified, fourteen are classified as WTO-plus and thirty-eight are WTO-extra. Aggregate depth thus varies from 1 to 52. We use this to calculate the average annual Depth of integration for each country in our sample. That is, for each country I in year t, Depth_it is the average of the total depth of all the agreements it belongs to in that particular year. As Figure A1 in the appendix shows, average depth for the countries in our sample varies from 15 for Bolivia and Ecuador to 30 for Haiti.Footnote ¹³

Our other main variable of interest is State Capacity_it of the host country i in year t. Significant debate has surrounded the conceptualization and operationalization of this term. Here, we follow Mann’s (Reference Mann1984, 189) understanding of state capacity as the ability of the state “to implement logistically political decisions through the realm.” Moreover, like Cárdenas (Reference Cárdenas2010), we understand state capacity to include not only aspects of bureaucratic efficiency but also the state’s ability to raise revenue from society, to deliver basic public goods, and to provide a sound legal and institutional framework that protects property rights. We thus use three indicators drawn from the World Bank’s World Governance Indicators (WGI), which seek to capture “the capacity of the government to effectively formulate and implement sound policies” (Reference Kaufmann, Kraay and MastruzziKaufmann, Kraay, and Mastruzzi 2010, 4). Government effectiveness measures the quality of public services, the quality and independence of the civil service, and the quality and credibility of policy formulation and implementation. Rule of law captures another aspect of state capacity, namely, the “quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence” (Kaufman, Kraay, and Mastruzzi 2010, 4). Finally, Regulatory quality measures the ability of the government to formulate and implement sound policies and regulations that promote private sector development. The three indices take measures of –2.5 to 2.5, with higher values indicating stronger performance. Even though they are not without criticism,Footnote ¹⁴ these indicators capture important dimensions of the regulatory and administrative capacities that, as discussed in the previous section, are crucial for the implementation of trade agreements. While the Rule of law variable measures the institutional features that we associate with the signaling mechanism, the other two variables could be viewed as more relevant to testing the implementation channel through which state capacity may condition the effectiveness of PTAs.

Our hypotheses concern the interactive effect of PTAs and state capacity on FDI inflows. We predict that the greater a country’s state capacity, the more likely it will be to implement its PTAs, or to signal credibility to investors, and hence the greater the FDI flows it will attract. To test these claims, we interact PTAs and Depth with each of the indicators of state capacity (in separate models, due to high correlation among them). We expect the coefficient of these interaction terms to be positive. We then consider the effect of Depth in trade agreements and its interaction with state capacity indicators.

In addition, we control for three factors that affect FDI inflows and are also associated with the number of agreements a country signs and with the level of state capacity, namely, level of economic development, measured as the log of the country’s per capita gross domestic product (GDP; in current US dollars) and market size, measured as the log of population, and the member country’s level of trade openness, measured as the log of the sum of its imports and exports of goods and services as a percentage of GDP. Data for these variables comes from the World Bank’s World Development Indicators.

The main empirical challenge in this research design is the possibility of endogeneity. Endogeneity may arise because of reverse causality, as an increase in FDI inflows might make countries more likely to sign PTAs. For example, a country that is host to multinational firms may face pressures to engage in preferential trade relations to expand the market opportunities and thus reduce trading costs of these investors. Similarly, it could be argued that higher FDI inflows contribute to building domestic institutional capacity. One way to deal with this problem is to rely on an instrumental variable approach. However, to the best of our knowledge, the literature has not identified convincing instruments for PTA formation, that is, satisfying both the relevance restriction (that the instrument be correlated with the causal variable—PTAs in this case) and the exclusion restriction (that the instrument is uncorrelated with any other determinants of FDI). To mitigate endogeneity concerns, we examine the lagged effects of PTAs and state capacity on FDI. By lagging the PTA variables —number and depth of PTAs—we substantially reduce the scope for endogeneity, while also allowing a time-adjustment for agents to become aware of PTAs and make their investment decisions.Footnote ¹⁵

Another endogeneity concern lies in the existence of omitted variables that may be correlated both with PTA formation and with FDI attraction. In this respect, we follow Medvedev (Reference Medvedev2012, 53) in assuming that the use of panel data with fixed effects “adequately addresses the endogeneity problem as long as the unobservables determining the likelihood of forming a PTA are not time-dependent.” Year dummies also help reduce the scope of omitted variables that may vary over time and affect all countries alike, such as technological change reducing transport costs or improving information technology connectivity that increases FDI and the propensity of countries to sign PTAs.Footnote ¹⁶

Results

Table 1 contains the results of our main regression analysis. We begin by estimating the individual effects of PTAs on FDI inflows using country and year fixed effects. The models in this table use the two-year lag of the annual count of PTAs, that is, we examine whether the number of PTAs a country belonged to in year t – 2 is associated with the level of foreign investment inflows in t. In model 1, the estimated coefficient for PTAs is negative but statistically insignificant. By contrast, the estimated coefficient for Depth is positive and statistically significant at the 1 percent level, suggesting that while the net effect of the number of trade agreements signed by a country does not significantly affect the level of FDI inflows, the depth of these agreements does.

Table 1 Number of PTAs, state capacity, and FDI inflows: Fixed effects models.

We then explore our first hypothesis, namely, whether the ability of countries to benefit from the PTAs they sign depends on the quality of their institutions. In model 2, the coefficient for PTAs × Rule of law is positive and statistically significant, in line with our expectations. However, the coefficient for the individual effect of PTAs, which maintains its negative sign, now becomes statistically significant.Footnote ¹⁷ We obtain similar results in models 3 when adding a two-year lag of the dependent variable (log of FDI inflows). However, a problem with this specification is that the lagged dependent variable is endogenous to the fixed effects in the error term, giving rise to dynamic panel bias (Reference RoodmanRoodman 2006). To overcome this, in models 4–6, we use the Blundell-Bond estimator.Footnote ¹⁸ In model 4, the estimated coefficient for PTAs × Rule of law is significantly positive, while the individual effect of PTAs remains significant and negative. Models 4 and 5 replace Rule of law with Government effectiveness and Regulatory quality, respectively, and while the interactive terms are positive, they are slightly below conventional levels of statistical significance.

These results are in line with H₁, namely, that the effect of the cumulative number of PTAs is conditional on the quality of members’ domestic institutions. They show, first, that signing more trade agreements does not necessarily translate into higher FDI inflows and may even decrease them. This is consistent with PTAs having a spaghetti-bowl effect by which the trade diversion effects offsets the expected trade creation effects for the signatory country and make investors indifferent about investing in the country. Yet while signing multiple agreements may work to deter FDI inflows, this negative effect is at least partially offset in the context of high performing domestic institutions. However, the fact that only the interaction of PTAs with Rule of law is statistically significant may suggest that the stability of business regulations and the effectiveness of government agencies are not as relevant as the strength of the rule of law in mediating the effects of trade agreements on FDI.

Figure 3 shows how the marginal effect of PTAs varies across observed levels of Rule of law—in other words, how the predicted impact of an additional PTA on FDI inflows increases with the rule of law score in the host country. For countries with very poor-quality rule of law, signing another agreement has a statistically significant negative effect on FDI. However, as rule of law improves this negative impact becomes smaller. In other words, the quality of domestic institutions somewhat mitigates the spaghetti-bowl effect of multiple agreements. Indeed, at values of Rule of law approximating 2, the marginal effect turns positive, although it is not statistically significant.

Figure 3 Average marginal effect of number of PTAs conditional on rule of law. Based on model 2, Table 1 (95% CI).

More agreements are therefore not necessarily a good thing for countries in the Americas. The results in Table 1, however, show a highly significant and positive estimated coefficient for Depth across specifications. All else equal, the deeper the agreements that a county belongs to, the higher the FDI inflows it attracts. Are these positive effects of depth also conditional on the level of state capacity and institutional quality of members? To test H₂, in the models presented in Table 2, we interact our indicators of state capacity with Depth. Models 1 and 4 show that both Depth and Depth × Rule of law have positive and statistically significant coefficients, even when controlling for past levels of FDI inflows (and using system generalized method-of-moments, or GMM, estimators). Moreover, the linear combination of these estimated coefficients, that is, the average marginal effect of Depth on FDI inflows, is also positive and statistically significant at the 1 percent level. All else equal, deeper integration increases FDI inflows, yet the magnitude of the effect depends on the quality of the rule of law in the host country. Similar findings are obtained when using Government effectiveness (models 2 and 5). The estimated coefficients for Depth and for the interaction terms with Government effectiveness are positive and statistically significant, even when the lagged dependent variable is not included in the model). Finally, the interaction of Depth and Regulatory quality is also positive and significant at the 10 percent level in the dynamic panel (model 6).

Table 2 PTA depth, state capacity, and FDI: Fixed-effects models.

Note: All models include country and year fixed effects. Robust standard errors in parentheses.

*** p < 0.01; ** p < 0.05; * p < 0.1.

To shed further light on the conditional relationship between Depth and institutional capacity, we examine the estimated marginal effect of this variable over a range of meaningful values Government effectiveness. As Figure 4 shows, the conditional relationship is not significant over very low values of this indicator. For Government effectiveness greater than –0.5, however, the marginal effect is positive and statistically significant. As the positive slope of the line shows, the higher the level of state capacity, the stronger the effect of Depth on FDI inflows. Similarly, Depth has a statistically significant effect on FDI inflows only for levels of Rule of law greater than 0, and this effect grows in magnitude as values for Rule of law.Footnote ¹⁹

Figure 4 Average marginal effect of depth conditional on governmental effectiveness. Based on model 2, Table 2.

Figure 5, in turn, shows the predicted FDI inflows at varying depths of integration for two medium-sized countries, such as for example Chile and Peru. Although both countries are similar in terms of GDP, which is sometimes used as an indicator of market size, Chile’s state capacity indicators are higher than Peru’s. For example, in 2014, the Rule of law score for Chile reached 1.43 while Peru’s was –0.52. If Peru, which in 2014 had fourteen PTAs, were to expand its Depth score (the average depth of all its trade agreements) from 25 to 26, it could expect to see FDI inflows increase by 0.2 percent. Instead, a country with Chile’s much stronger rule of law institutions would derive much higher benefits from deepening integration. An increase in Depth from 22 to 23 is associated with an increase in predicted FDI inflows by a factor of 1.19 (i.e., 19 percent), all other factors being held constant.

Figure 5 PTA depth and FDI inflows conditional on rule of law. Based on model 1, Table 2.

In sum, the results in Tables 1 and 2 provide support for our central claim that the benefits of trade agreements are conditional on the quality of domestic institutions. The strength of the rule of law, the effectiveness of the government in formulating and implementing policies, and the quality of the regulatory process, seem to condition the effects of both the number of PTAs signed and their depth on FDI levels. In particular, our findings give support to our claim that very poor levels of state capacity, by hindering implementation and credibility, undermine the effectiveness of PTAs in attracting FDI. Far from increasing their attractiveness to foreign investors, countries with very low state capacity may undermine it when signing more and deeper trade agreements.

Robustness checks

Our results are robust to alternative estimation techniques and different measures of the dependent and independent variables. We first estimate our baseline models using FDI as a percentage of GDP as the dependent variable (Table A3 in the online appendix). Although the estimated coefficient for the interaction of state capacity and the cumulative number of PTAs is positive, it is not statistically significant (model 1). However, we still find support for our second hypothesis, concerning the interaction of Depth and state capacity. The interactions of Depth and both Rule of law and Government effectiveness remain positive and statistically significant (models 2–5). Results are unchanged when using the system GMM model and including the two-year lag of the dependent variable (FDI/GDP). These findings are to some extent consistent with our main findings, which suggest that depth has a clearer effect than quantity of agreements on FDI.

In an attempt to look at whether participation in PTAs and state capacity influence not just the amount but also the concentration of FDI, we run the model using the number of announced investment projects into each country as the dependent variable.Footnote ²⁰ The results show that while the number of PTAs does not significantly affect the count of investments announced for each country, the average depth of these agreements is negatively associated with it (Table A5). In other words, the deeper the PTAs a country belongs to, the fewer the investment projects that it attracts. The interaction of Depth and state capacity, however, fails to reach statistical significance. These results, while less robust than our central findings, suggest that deeper agreements may lead to higher but also more concentrated FDI inflows—that is, fewer but larger projects.

We then replace WGI indicators of state capacity with an indicator of Quality of government developed by Teorell et al. (Reference Teorell, Dahlberg, Holmberg, Rothstein, Pachón and Svensson2018) with data from International Country Risk Guide.Footnote ²¹ The interaction of the number of PTAs and Quality of government indicator is positive and statistically significant, providing further support for our first hypothesis (Table A5). In line with H₂, Depth × Quality of government is also statistically significant and has the expected sign. Similar results are obtained with the World Bank’s Country Policy and Institutional Assessment indicator Quality of public administration.Footnote ²²

The article’s main findings are also robust to the exclusion of specific countries—Brazil, Argentina, Mexico, and Chile—and regional groupings. We find support for both H₁ when excluding all the main trade blocs in the Americas, namely MERCOSUR, CAN, CACM, and CARICOM. However, our findings concerning the interaction of Depth of integration and state capacity appear to be sensitive to the exclusion of CARICOM countries (Table A6). Finally, we try including a different structure of lags for PTAs and Depth. The central findings remain largely unchanged when using the three- and the five-year lag of these two variables. However, when including a one-year lag and no lags, the estimated coefficients for the interaction terms of Depth and the state capacity indicators are not statistically significant, suggesting that depth of integration takes at least a couple of years to influence FDI (Table A7).