Deep Trade Agreements and FDI in Partial and General Equilibrium: A Structural Estimation Framework

Abstract

This paper quantifies the relationships between deep trade agreements and foreign direct investment (FDI). The analysis relies on a structural framework that simultaneously enables (a) estimating the direct impact of deep trade agreements on FDI, (b) translating the partial deep trade agreement estimates into general equilibrium effects on FDI, and (c) obtaining partial deep trade agreement effects on trade and quantifying the impact of deep trade agreements on FDI through trade. The effects of deep trade agreements on both trade and FDI are sizeable, positive, and statistically significant. A counterfactual analysis suggests that together with direct and indirect channels deep trade agreements have contributed to a large but asymmetric increase in inward versus outward FDI.

1. Introduction

Most modern preferential trade agreements (PTAs) include a variety of investment provisions. As pointed out by Crawford and Kotschwar (2020), “Following the entry into force of NAFTA and the GATS, trade negotiators increasingly began to incorporate into PTAs a broad set of investment provisions that liberalize, protect, and regulate investments” (p. 145). The increase, both in absolute and in relative terms, in the number of PTAs with investment provisions is depicted in fig. 1, which comes from Crawford and Kotschwar (2020).

Using the World Bank’s Database on the Content of Regional Trade Agreements (DCRTA), cf. Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020), we complement fig. 1 by plotting the number of country pairs that have signed a trade agreement that includes investment provisions in absolute terms (in the left-hand panel of fig. 2) and relative to all country pairs with PTAs (in the right-hand panel of fig. 2). This figure corroborates the evidence from fig. 1 by depicting a remarkable increase in the country pairs that have negotiated investment together with trade, especially since the early 1990s, as noted in the opening quote from Crawford and Kotschwar (2020).

Despite the increase in the number and importance of investment provisions in the negotiations and implementation of PTAs, there is relatively little and mixed evidence on the effectiveness of such provisions in promoting foreign direct investment (FDI). For example, the authoritative surveys of Eicher, Helfman, and Lenkoski (2012) and Blonigen and Piger (2014) on the determinants of FDI do not account for such provisions.¹ They do account for bilateral investment treaties (BITs) that have traditionally regulated foreign investments, but the empirical evidence of the effectiveness of BITs is mixed (see Lesher and Miroudot (2006) and Laget, Rocha, and Varela (2021) for examples). Nowadays, more and more trade agreements include investment provisions, which may partly explain the decline of the role of BITs. However, only very recently, some papers (e.g., Kox and Rojas-Romagosa 2020; Laget, Rocha, and Varela 2021) have studied the impact of deep trade agreements (DTAs) and various PTA provisions (disciplines) on FDI, offering mixed evidence of the effectiveness of investment provisions.²

Against this backdrop, we make three contributions to the existing literature on the links between deep trade liberalization and FDI. First, we contribute to the debate on whether deep trade agreements with investment provisions stimulate FDI by estimating the direct/partial equilibrium effects of DTAs and DTAs with investment and other provisions on FDI. While traditional trade agreements liberalize trade, DTAs cover additional policy areas, such as international flows of investment, international flows of labor, the protection of intellectual property rights, and the environment. Hence, the objective of the DTAs is to go beyond trade and to lead to deeper integration between the agreement members.

Besides capturing the movement of goods, services, capital, people, and ideas, DTAs also include enforcement provisions limiting the discretion of firms and governments in the importing and exporting country (see Mattoo, Rocha, and Ruta (2020) for an overview and discussion of typical areas of DTAs). Investment provisions are one important area of DTAs that go beyond trade. Crawford and Kotschwar (2020) distinguish five categories of investment provisions that are present in many trade agreements: (a) definitions and scope; (b) investment liberalization; (c) investment protection; (d) social and regulatory goals; and (e) institutional aspects and dispute settlement. For our DTA variable, we use the variable “pta_mapped” from Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020), which is 1 whenever the depth variable is greater than 0, and 0 otherwise. The depth variable counts the number of provisions included in a PTA.

Second, we use our partial estimates to obtain novel general equilibrium (GE) estimates of the effects of DTAs on FDI. Third, within the same structural framework, we obtain estimates of the effects of DTAs on trade flows, and we translate those effects into general equilibrium effects of DTAs on FDI through trade liberalization.

Guided by the theoretical model of Anderson, Larch, and Yotov (2019),³ we specify two estimating gravity equations—one for trade and one for FDI, which are (a) consistent with and representative of a large number studies that quantify the impact of various determinants on FDI (e.g., Eicher, Helfman, and Lenkoski 2012; Blonigen and Piger 2014; Kox and Rojas-Romagosa 2020; Laget, Rocha, and Varela 2021), and (b) capitalize on the latest developments in the trade gravity literature (e.g., Head and Mayer 2014; Yotov et al. 2016). Specifically, we rely on the Poisson pseudo-maximum-likelihood estimator to account for potential heteroskedasticity in the bilateral trade and FDI data and to take advantage of the information in the zero trade and FDI flows (cf. Santos Silva and Tenreyro 2006, 2011). In addition, we employ a rich set of fixed effects (including origin time, destination time, and directional country-pair fixed effects), which control for and absorb all possible country-specific and time-invariant bilateral determinants of trade and FDI.

We have potential concerns that our key variables of interest, PTAs and DTAs, are endogenous. There is a comparably large literature dealing with the potential endogeneity of trade policy variables in general and trade agreements to be specific.⁴ As it is hard to come up with good instruments, one of the most followed suggestions in the literature by Baier and Bergstrand (2007) is to include country-pair fixed effects to control for all non-time-varying bilateral effects. We follow this literature to mitigate potential endogeneity concerns, acknowledging that some endogeneity concerns may still be left if time-varying country-pair factors that are not accounted for in our model are correlated with PTA and DTA.⁵ In addition to PTAs and DTAs, we control other policy variables such as WTO membership, economic sanctions, and bilateral investment treaties.

To perform the empirical analysis we build a balanced panel data set for 89 countries covering more than 96 percent of world GDP and more than 94 percent of FDI throughout the sample period, 1990–2011. Our data set covers foreign direct investment, trade agreements, trade flows, gross domestic product (GDP), employment, physical capital, bilateral investment treaties, sanctions, and WTO membership. An important feature of the data set is that we capitalize on the richness of the Database on the Content of Regional Trade Agreements (DCRTA), cf. Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020). Specifically, the DCRTA enables us to distinguish between several indicator and continuous PTA variables, including a standard dummy variable for PTAs, an indicator variable for DTAs, an indicator for DTAs that include investment provisions, and two continuous variables for the overall depth of DTAs and the depth of DTAs with investment provisions.

Three main findings stand out from our estimates of the effects of DTAs on trade. First, we find an average non-significant impact of PTAs on trade in our sample. However, second, we obtain positive and statistically significant estimates of the effects of deep trade agreements. Specifically, our estimates suggest that the DTAs in our sample have led to a 14.888 percent (std.err. 2.856) increase in bilateral trade among member countries. Finally, our estimates reveal that deeper trade agreements (as measured by the number of provisions) lead to larger increases in the trade flows among DTA members. Depending on the number of provisions that they include, the DTAs in our sample have led to trade increases between 0.784 percent (std.err. 0.300) and 32.472 percent (std.err. 14.194). Overall, our estimates of the DTA effects in trade are consistent with findings from recent studies that have utilized the database on the Content of Regional Trade Agreements and reinforce the view that “depth” matters for the effectiveness of PTAs.⁶

Similar to our results for trade, the estimates of the effects of DTAs on FDI are also heterogeneous. Specifically, we do not obtain significant estimates of the effects of PTAs and DTAs on FDI. However, when we zoom in on the effects of DTAs that include investment provisions, we obtain a positive, sizable, and statistically significant estimate, which suggests that, on average, the PTAs with investment provisions in our sample have led to a 21.038 percent (std.err. 11.513) increase in FDI between their members. This result is consistent with earlier findings from Lesher and Miroudot (2006). We also obtain positive estimates of the effects on FDI of several other DTA provisions including labor-market regulations, export taxes, public procurement, and state-owned enterprises. This analysis reinforces and complements the findings of Laget, Rocha, and Varela (2021) who study the impact of different DTA provisions with firm-level data for the years 2003–2015. Finally, our estimates do not reveal a significant impact of the increase in the depth (number of provisions) on FDI. Our results suggest that an increase in the number/complexity of some investment provisions (e.g., related to transparency and to regulations) may decrease FDI.

We use the structural model in combination with our estimates of the partial effects of DTAs on trade and FDI to quantify the GE impact of DTAs on FDI. As discussed in more detail in A caveat with our GE analysis is that the underlying theory is based on the assumption of non-rival technology FDI, while our data include all/aggregate FDI flows.⁷ We focus the analysis on inward usage of technology FDI per country and outward technology FDI stocks per country used abroad.

The main conclusions from this analysis are as follows. DTAs have had large and strongly asymmetric effects on FDI. The DTAs that were in force in 2011 have contributed to about 2 percent of inward FDI in the world and about 43 percent of outward FDI. The large average effect of outward FDI is driven by some large outward FDI countries (such as China and the United States), where, consistent with our theoretical model of non-rival capital, any change in the technology stock of these countries has a multiplying effect due to the usage in many other countries, resulting in a large boost in outward FDI stock, i.e., in the usage of a country’s technology capital abroad. We view our result about the disproportionately large impact of outward FDI as novel and potentially important from a policy perspective, both for the negotiations of trade and investment agreements and for properly quantifying their implications.

Finally, we also find that changes in trade costs due to DTAs have led to additional boosts in FDI through the GE links between trade and FDI in our model. Specifically, through their impact on trade costs, the 2011 DTAs in our model have boosted inward FDI by an additional 1 percentage point and outward FDI by an additional 11 percentage points, i.e., effects that are about a quarter of the corresponding estimates due to FDI liberalization. By demonstrating that the impact of DTAs on FDI through trade is significant, we complement some recent work on the GE links between DTAs and trade, cf. Fontagné et al. (2023), and also, from a broader perspective, papers that have studied the GE links between trade liberalization and FDI, cf. Baltagi, Egger, and Pfaffermayr (2008), Tintelnot (2017), and Anderson, Larch, and Yotov (2019).

The rest of the paper is organized as follows. First, we presents the methodological foundations of our analysis, and the specifications of our estimating equations for bilateral trade flows and FDI. Next, we describe the main variables and the corresponding data sources that we use to construct them, followed by a presentation and discussion our partial estimates and our GE results. At the end, we conclude and offer a supplementary online appendix that includes the derivations of the theoretical model, offers some additional estimation results and analysis of the effects of DTAs on trade and FDI, and includes additional counterfactual results and analysis.

2. Methods

To quantify the impact of deep trade agreements on FDI, we rely on the theoretical framework of Anderson, Larch, and Yotov (2019). While our current contribution is purely empirical, we find it helpful to summarize the model of Anderson, Larch, and Yotov (2019) for two reasons. First, it will motivate the two key estimating equations for trade and FDI, Second, we will use the model to perform counterfactual analysis.

2.1. Theoretical Foundations

Anderson, Larch, and Yotov (2019) derive the following steady-state structural system that describes the relationships between trade, domestic investment, and FDI:⁸

Equations (1)–(4) may look familiar because they represent the structural gravity trade system of Anderson and van Wincoop (2003). Equation (1) is the standard structural gravity equation, which captures the fact that the exports from source i to destination j, |$X_{ij}$|⁠, are proportional to the sizes of the two countries, i.e., the exporter’s output, |$Y_i$|⁠, and the importer’s expenditure, |$E_j$|⁠, and inversely proportional to the trade costs between the two countries, including the direct bilateral trade frictions, |$t_{ij}$|⁠, and the GE multilateral resistances. The multilateral resistance terms (MRTs, outward and inward, respectively), denoted by |$\Pi _{i,t}^{1-\sigma }$| and |$P_{j,t}^{1-\sigma }$|⁠, consistently aggregate bilateral trade costs and decompose their incidence on the producers and the consumers in each region.

Equations (5) and (6) define the value of production and the expenditure in country j, respectively. Specifically, equation (5) is the production value function. Here, |$p_{j,t}$| denotes the factory-gate price of good (country) j at time t, |$A_{j,t}$| is the local, country-specific technology, |$L_{j,t}$| is country-specific (internationally immobile) labor, and all other variables are defined earlier. The last term, |$\prod _{i=1}^{N}(\max \lbrace 1,\omega _{ij,t}M_{i,t}\rbrace )^{\eta _{i}}$|⁠, is the global technology stock in j, which may include non-rival technology capital, |$M_{i,t}$|⁠, from all other countries⁹ which is used in j subject to investment frictions (⁠|$1 \ge \omega _{ij,t} \ge 0$|⁠).¹⁰ By assuming |$\sum _{i=1}^{N}\eta _{i}=1$|⁠, we impose constant returns to scale. Finally, the |$\max$|-function implements the notion that there is some world knowledge of technology capital freely available to all countries and ensures that there is always some technology capital available for all countries. Equation (6) defines expenditure as the sum of total nominal output |$(Y_{j,t})$| plus rents from foreign investments, |$\phi \eta _{j}\sum _{i \in \mathbb {N}_{ji,t}}Y_{i,t}$|⁠, minus rents accruing to foreign investments, |$\phi Y_{j,t}\sum _{i\in \mathbb {N}_{ij,t}}\eta _{i}$|⁠.

Equation (7) is the solution for physical capital, where |$\Psi _j$| is a composite model parameter. Intuitively, the direct relationship between |$K_{j}$| and |$Y_{j}$| reflects the fact that there will be more investment the higher the value of the marginal product of physical capital. The inverse relationship between |$K_{j}$| and |$P_{j}$| can be interpreted through the lens of the law of demand, i.e., if |$P_{j}$| is interpreted as the price of investment goods. Alternatively, if |$P_{j}$| is the price of consumption or technology goods, then the intuition for the inverse relationship is that there will be less investment when the opportunity cost of it (i.e., investment in consumption or technology goods) is higher.

Finally, equation (8) is the structural gravity equation for FDI, where |$\mathrm{FDI}_{ij}$| is the value of the stock of FDI from origin i at destination j, |$\Gamma _i$| is a composite country-specific constant term, and all other variables are defined above. Intuitively, and similar to the gravity model of trade, (8) captures the direct relationship between FDI and the sizes of the source and the destination countries. The explanation for the inverse relationship between FDI and |$P_{j}$| is similar to the relationship between physical capital and |$P_{j}$|⁠. The inverse relationship between FDI and |$M_{i}$| is a reflection of the law of diminishing marginal productivity, i.e., the larger the stock of technology capital in country i, the smaller the marginal productivity by an additional unit of investment in technology.¹¹

For given parameters and variables that are exogenous in the model, we can use system (1)–(8) to simulate the impact of deep trade liberalization on trade and investment in the world. We capitalize on this in our counterfactual analysis.¹² Next, we rely on system (1)–(8) to specify the econometric models that will deliver our key estimates of the direct impact of DTAs on trade and FDI.

2.2. From Theory to Empirics

A key objective and contribution of this paper is to test for links between DTAs, trade, and FDI. Establishing such links, and obtaining estimates of the corresponding direct/partial effects of DTAs on trade and FDI, would also enable us to translate them into GE effects of trade liberalization on FDI through the structural links that we just described in the previous section. In this section, we rely on system (1)–(8) to specify our estimating equations for trade and FDI. Specifically, as noted earlier, equation (1) is the standard structural gravity equation from the trade literature, while (8) is our theoretical gravity equation for FDI. To estimate both equations, we will capitalize on the latest developments in the empirical trade literature.¹³ We start with the estimating equation for bilateral trade flows:

Here, |$X_{ij,t}$| denotes nominal (cf. Baldwin and Taglioni 2006) exports from i to j at time t. Consistent with theory, |$X_{ij,t}$| includes international and domestic trade flows (cf. Yotov 2022). Estimating equation (9) includes three sets of fixed effects. The terms |$\psi _{i,t}$| and |$\phi _{j,t}$| denote exporter-time and importer-time fixed effects, respectively, which will account for the country-size and the multilateral resistance terms (cf. Anderson and van Wincoop 2003) in equation (1), and also for any other observable or unobservable factors that affect trade flows on the exporter or on the importer side, including input-output links and other network and production relationships (possibly involving FDI), which may not even be captured by our theoretical model. The term |$\mu _{ij}$| denotes a set of pair fixed effects, which will control for all time-invariant bilateral trade costs (cf. Egger and Nigai 2015) and will mitigate endogeneity concerns of the bilateral policy variables in our setting (cf. Baier and Bergstrand 2007), including DTAs. The pair fixed effects will also fully absorb the effects of all policies concluded before the beginning of our period of investigation. Our main results will be obtained with directional pair fixed effects, which allow for asymmetric time-invariant trade costs depending on the direction of trade flows, i.e., from i to j vs. from j to i.

The vector |$\mathbf {GRAV\_TRADE}_{ij,t}$| includes a set of time-varying bilateral control variables that control for WTO membership (⁠|$\mathrm{WTO}_{ij,t}$|⁠), economic sanctions (⁠|$\mathrm{SANCT}_{ij,t}$|⁠), and bilateral investment treaties (⁠|$\mathrm{BIT}_{ij,t}$|⁠). Even though, by definition, BITs are not designed to promote trade, we include the BIT dummy in our trade model for three reasons. First, we do this for symmetry between our trade and FDI estimating equations. An alternative motivation could be based on some theories of substitutability, i.e., making trade easier could decrease FDI and vice versa. Finally, by promoting FDI, the BITs may stimulate intra-firm trade. Our data do not allow us to test these hypotheses directly. However, as will be seen shortly, in the results section, BITs have a strong impact on trade.

In addition, we also include a full set of time-varying border indicators (⁠|$\sum _t \mathrm{BRDR}_{ij,t}$|⁠), which take the value 1 for international trade in a given year and 0 for domestic trade flows and, therefore, capture any common globalization trends (e.g., improvements in communication, transportation, and communication). Finally, the vector |$\mathbf {DTA\_TRADE}_{ij,t}$| includes the variables whose estimates would be of central interest to us. Specifically, we will differentiate between the effects of preferential trade agreements (⁠|$\mathrm{PTA}_{ij,t}$|⁠), the effects of deep trade agreements (⁠|$\mathrm{DTA}_{ij,t}$|⁠), and we will allow for the effects of DTAs to vary depending on their depth (⁠|$\mathrm{DEPTH}_{ij,t}$|⁠), which will be measured by the number of provisions that they include. Specifically, we utilize the World Bank’s database on deep trade agreements (DTA) (Hofmann, Osnago, and Ruta 2019; Mattoo, Rocha, and Ruta 2020), which provides information on the presence of any trade agreement between i and j at time t (⁠|$\mathrm{PTA}_{ij,t}$|⁠) and on the number of provisions in each trade agreement. If the number of provisions in a trade agreement is greater than zero, then it is classified as a deep trade agreement (⁠|$\mathrm{DTA}_{ij,t}$|⁠).

We estimate equation (9) with the Poisson pseudo-maximum-likelihood (PPML) estimator to account for the presence of heteroskedasticity in the trade data and to take advantage of the information contained in the zero trade flows; cf. Santos Silva and Tenreyro (2006, 2011). Despite the tradition in the trade literature to use interval data for gravity estimations, we follow the recent recommendations of Egger, Larch, and Yotov (2022) and employ all available data for consecutive years.¹⁴ Finally, we cluster the standard errors by country pair.

Next, guided by equation (8), we specify our estimating gravity equation for FDI as follows:

Here, |$\mathrm{FDI}_{ij,t}^{\mathrm{value}}$| is the value of FDI stock from origin i to destination j at time t. Capitalizing on the developments in the bilateral trade and FDI literature, and for consistency with our estimating equation for trade flows, we specify our FDI econometric model to be as close as possible to our estimating equation for trade flows given in equation (9). Specifically, we use the same estimator (i.e., PPML), we include the same set of fixed effects (i.e., origin-time fixed effects (⁠|$\psi _{i,t}$|⁠), destination-time fixed effects (⁠|$\phi _{j,t}$|⁠), and directional pair fixed effects (⁠|$\mu _{ij}$|⁠)), and we employ the same set of time-varying policy covariates (i.e., indicators for WTO membership (⁠|$\mathrm{WTO}_{ij,t}$|⁠), for bilateral investment treaties (⁠|$\mathrm{BIT}_{ij,t}$|⁠), and for sanctions (⁠|$\mathrm{SANCT}_{ij,t}$|⁠)). Finally, just as in our trade specification, we follow the recommendation of Egger, Larch, and Yotov (2022) to use consecutive-year data, and we use the same clustering (i.e., by country pair).

Even though, from an econometric perspective, we will use the same set of exporter-time and importer-time fixed effects as in our trade equation, the country-time fixed effects in the FDI model would proxy and account for different variables. Following the existing empirical FDI literature,¹⁵ possible robust determinants of FDI in the country of origin include corporate tax rate, corruption, and bureaucratic red tape, while possible candidates at the destination include level of corruption, internal tensions, corporate tax rate, bureaucratic red tape, and quality of institutions.¹⁶ Finally, the pair fixed effects in (10) will absorb and account for bilateral distance, common official language, and colonial relationships, which, similar to the trade literature, are among the most robust FDI determinants in Eicher, Helfman, and Lenkoski (2012) and Blonigen and Piger (2014). Using pair fixed effects also mitigates concerns regarding our inability to distinguish between horizontal and vertical FDI.

There are two differences between equations (9) and (10). First, we cannot include the set of time-varying border effects (⁠|$\sum _t \mathrm{BRDR}_{ij,t}$|⁠) in equation (10) since we only use data on international transactions. This is why we use different notation for the vector of time-varying gravity covariates (⁠|$\mathbf {GRAV\_FDI}_{ij,t}$|⁠). We also allow for potential differences in the estimated impact of the common policy covariates by denoting the vector of their estimates |$\widetilde{{\boldsymbol \alpha }}$|⁠. Second, and more importantly for our purposes, we use a different set of variables to capture the impact of DTAs on FDI in the vector |$\mathbf {DTA\_FDI}_{ij,t}$|⁠. Specifically, in addition to including indicators for PTAs (⁠|$\mathrm{PTA}_{ij,t}$|⁠) and DTAs (⁠|$\mathrm{DTA}_{ij,t}$|⁠), we add two more covariates. First, motivated by Osnago, Rocha, and Ruta (2019), Crawford and Kotschwar (2020), and Laget, Rocha, and Varela (2021), we include a separate indicator variable (⁠|$\mathrm{INV}_{ij,t}$|⁠) that takes a value of 1 for agreements that include investment provisions. Second, we also account for depth of the investment treaties by using a count variable (⁠|$\mathrm{INV\_DEPTH}_{ij,t}$|⁠) for the number of investment provisions within the agreements with investment provisions, i.e., similar to the relationship between |$\mathrm{PTA}_{ij,t}$| and |$\mathrm{DEPTH}_{ij,t}$| on the trade side, |$\mathrm{INV\_DEPTH}_{ij,t}$| is a continuous variable that is equal to 0 when |$\mathrm{INV}_{ij,t}$| is 0.¹⁷

Estimating equations (9) and (10) will deliver the estimates of the effects of DTAs on trade and investment that we will describe when we discuss our estimation results and use to obtain GE results. Before that, we describe our data.

3. Data and Sources

To perform the empirical analysis we build a balanced panel data set for 89 countries for the years 1990–2011, covering more than 96 percent of world GDP and more than 94 percent of FDI throughout the sample period.¹⁸ Our data set includes the following variables: foreign direct investment, trade agreements, trade flows, gross domestic product (GDP), employment, physical capital, bilateral investment treaties, sanctions, and WTO membership. Next we describe, in turn, the sources to obtain these variables, as well as their construction.

• FDI Data. We use two sources to construct the FDI variable, (⁠|$\mathrm{FDI}_{ij,t}$|⁠), which takes a central stage in our analysis. The main source for FDI data is the Bilateral FDI Statistics database of the United Nations Conference on Trade and Development (UNCTAD). These data can be accessed at https://unctad.org/topic/investment/investment-statistics-and-trends. UNCTAD’s FDI data cover inflows, outflows, inward stock, and outward stock for 206 countries over the years 1990–2011. Data are collected from national sources and international organizations and to ensure maximum coverage the data are mirrored. The second source of FDI data is the International Direct Investment Statistics database, constructed and maintained by the Organization for Economic Co-operation and Development (OECD). OECD’s data offer detailed statistics for inward and outward foreign direct investment flows and positions (stocks) of the OECD countries, including transactions between the OECD members and non-member countries. We use the OECD data to ensure consistency and maximum coverage. Finally, we note that, given our theory, we focus our analysis on FDI stocks (positions), which is also the FDI category for which most data are available.¹⁹

  Before we continue, we note that, despite its relatively long period and large number of countries covered, our data set has some limitations. Specifically, we cannot distinguish between vertical and horizontal FDI and between FDI in different sectors or across firms. These caveats imply that our results may be masking some important heterogeneity across the dimensions not covered, and our findings should be interpreted accordingly.

• Trade Agreements Data. To account for the presence and depth of trade agreements, we use the World Bank’s database on deep trade agreements (DTA); cf. Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020) (https://datatopics.worldbank.org/dta/about-the-project.html).²⁰ Capitalizing on the rich dimensionality of the DTA database, we construct and utilize several variables for our analysis. The variable |$\mathrm{PTA}_{ij,t}$| is an indicator for the presence of any trade agreement between i and j at time t. The variable |$\mathrm{DTA}_{ij,t}$| is an indicator denoting the presence of a deep agreement between i and j at time t. The variable |$\mathrm{DEPTH}_{ij,t}$| is a count variable for the number of provisions in the corresponding DTA between i and j. The variable |$\mathrm{INV}_{ij,t}$| is an indicator that takes a value of 1 if the DTA between i and j includes investment provisions. Finally, |$\mathrm{INV\_DEPTH}_{ij,t}$| is a count variable for the number of investment provisions in the corresponding DTA between i and j. For further details on the general features of the DTA database, we refer the reader to Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020). In addition, for analysis with a specific focus on investment provisions in the DTAs, we refer the reader to Crawford and Kotschwar (2020).

• Production Data. Data on GDP, employment, and capital stocks are from the Penn World Tables 8.0; cf. Feenstra, Inklaar, and Timmer (2013) (http://www.rug.nl/research/ggdc/data/pwt/). For data on GDP, we employ output-side real GDP at current PPPs (⁠|$\mathrm{CGDP}^{o}$|⁠), which compares relative productive capacity across countries at a single point in time, as the initial level in our counterfactual experiments, and we use real GDP using national-accounts growth rates (⁠|$\mathrm{CGDP}^{\mathrm{na}}$|⁠) for our income-based cross-country growth regressions. We measure employment in effective units by multiplying the number of persons engaged in the labor force with the Human Capital Index, which is based on average years of schooling. Finally, capital stocks in the Penn World Tables 8.0 are constructed based on accumulating and depreciating past investments using the perpetual inventory method.

• Trade Data. Data on international trade flows come from the United Nations Statistical Division (UNSD) Commodity Trade Statistics Database (COMTRADE). We complement the international trade flows data with data on domestic trade flows from Anderson, Larch, and Yotov (2020), which we use for both the estimation and the counterfactual analysis. Anderson, Larch, and Yotov (2020) construct domestic trade flows at the aggregate level in two steps. First, they use the ratio between aggregate manufacturing in gross values and total exports of manufacturing goods to construct a multiplier at the country-time level. (Data on gross manufacturing production came from the United Nations IndStat database.) Then, they use this multiplier along with data on aggregate exports to project the values for domestic sales. The availability of data on domestic trade flows predetermined the time coverage of our estimating sample.

• Other Data. Finally, we employ the following additional covariates as control variables: We control for the presence of bilateral investment treaties with an indicator variable |$\mathrm{BIT}_{ij,t}$|⁠, which comes from the UNCTAD’s data on international investment agreements (IIAs), which can be found at http://investmentpolicyhub.unctad.org/IIA. Data on sanctions come from the Global Sanctions Database (GSDB); cf. Felbermayr et al. (2020) and Kirilakha et al. (2021) (http://www.rug.nl/research/ggdc/data/pwt/). We use the GSDB to include an indicator variable (⁠|$\mathrm{SANCT}_{ij,t}$|⁠) for the presence of sanctions in our estimations. Finally, data on WTO membership, captured by an indicator variable |$\mathrm{WTO}_{ij,t}$| in our analysis, come from the Dynamic Gravity Dataset (DGD) of the US International Trade Commission; cf. Gurevich and Herman (2018) (http://www.rug.nl/research/ggdc/data/pwt/).

4. Empirical Findings and Analysis

We first present our partial estimates of the impact of DTAs on trade and FDI. Then, we translate the partial estimates into corresponding GE effects and analyze the total impact of DTAs on FDI within our framework.

4.1. Estimation Results

Our estimates of the effects of DTAs on trade are presented in table 1. All estimates are obtained with the PPML estimator with three-way fixed effects, including exporter-time, importer-time, and directional country-pair fixed effects. In addition, all specifications use time-varying border dummy variables to control for the presence of common globalization trends.

The estimates in column (1) include a single indicator variable, |$\mathrm{PTA}_{ij,t}$|⁠, that reflects the presence of a trade agreement of any type (e.g., deep or shallow) between i and j at time t. We note that, while positive, the estimate on |$\mathrm{PTA}_{ij,t}$| is economically small and not statistically significant. A possible explanation for this result is that we impose a common effect for all trade agreements in our sample, regardless of their type and depth. We demonstrate that this is indeed the case in column (3) of table 1. Before that, however, we add additional policy covariates in column (2) and discuss their estimates.

In column (2) of table 1, we add indicator variables for WTO membership (⁠|$\mathrm{WTO}_{ij,t}$|⁠), bilateral investment treaties (⁠|$\mathrm{BIT}_{ij,t}$|⁠), and economic sanctions (⁠|$\mathrm{SANCT}_{ij,t}$|⁠). First, we note that the estimate of the impact of WTO is positive, large, and statistically significant. This result is at odds with some of the existing literature, e.g., Rose (2004) and Esteve-Pérez, Gil-Pareja, and Llorca-Vivero (2020), who find that WTO membership did not promote international trade; however, our estimate on |$\mathrm{WTO}_{ij,t}$| confirms the findings of Larch et al. (2024) for positive WTO effects when domestic trade flows are used to estimate gravity equations. In robustness analysis, which appears in the supplementary online appendix, we confirm that when the model is estimated without domestic trade flows, the estimate of the effects of WTO is smaller and it is not statistically significant.

Second, we obtain a positive, sizable, and statistically significant estimate of the impact of bilateral investment treaties (BITs) on trade flows. A possible explanation for this result is multinational production. Third, we do not obtain a significant estimate of the impact of economic sanctions on trade. This result is consistent with estimates from Felbermayr et al. (2020), who argue that average estimates of the effects of sanctions may mask significant heterogeneity across the effects of sanctions by type. In addition, Kirilakha et al. (2021) demonstrate that the relative importance of trade sanctions has fallen significantly over time. In robustness analysis, we allowed for differential effects of different types of sanctions. This did not affect our main findings and conclusions.

In column (3) of table 1 we allow for heterogeneous effects between shallow and deep trade agreements. To this end, we capitalize on the data from World Bank’s DCRTA (cf. Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020)) to define |$\mathrm{DTA}_{ij,t}$| as an indicator that takes a value of 1 for deep trade agreements (i.e., we use the variable “pta_mapped” from the DCRTA database), and it is equal to 0 otherwise. Thus, by construction, the observations that take a value of 1 in the |$\mathrm{DTA}_{ij,t}$| variable are a subset of the observations that are equal to 1 in the |$\mathrm{PTA}_{ij,t}$| dummy from column (2). The main finding from column (3) of table 1 is encouraging and expected. Specifically, we obtain a positive and statistically significant estimate on |$\mathrm{DTA}_{ij,t}$|⁠, which suggests that, on average, the deep trade agreements in our sample have led to a 14.888 percent (std.err. 2.856) increase in bilateral trade among member countries relative to countries that only conclude a PTA.

Our positive and significant estimates of the impact of DTAs are consistent with and reinforce the general message from Fernandes, Rocha, and Ruta (2021) that DTAs have been effective in stimulating international trade. However, the DTA estimate that we obtain seems relatively small, which is consistent with a recent survey article (Larch and Yotov 2024), which traces the evolution of the estimates of trade agreements over the past sixty years. We investigate this puzzling finding further in the supplementary online appendix, where we show that using domestic trade flows and/or not accounting for common globalization effects leads to larger DTA estimates. We also show that the effects of the DTAs in the 1990s vs. 2000s are very similar and we obtain similarly small estimates with two alternative data sets. Finally, we show that the DTA estimates are very heterogeneous across disaggregated sectors.

In column (4) of table 1, we additionally split DTAs that contain investment provisions (⁠|$\mathrm{INV}_{ij,t}$|⁠). We find that deep trade agreements with investment provisions have larger effects on bilateral trade flows relative to DTAs without investment provisions. This is in line with our findings for BITs before and shows that investment provisions may spur trade potentially due to the importance of trade between headquarters and affiliates of multinational firms.

In column (5) of table 1, we use the DTA database to construct a continuous variable (⁠|$\mathrm{DEPTH}_{ij,t}$|⁠), which counts the number of provisions within each of the DTAs in our sample. The number of provisions across the DTAs in our sample varies between 12 and 432. The main result from column (5) is that, on average, the deeper the agreement, the more it would promote trade among its members. Specifically, we obtain a positive and statistically significant estimate on |$\mathrm{DEPTH}_{ij,t}$| (0.00065, std.err. 0.00024), which is consistent with the findings from Osnago, Rocha, and Ruta (2019). Our estimate suggests that depending on the number of provisions that they include, the DTAs in our sample have led to trade increases between 0.784 percent (std.err. 0.300) and 32.472 percent (std.err. 14.194). We capitalize on this variation in the counterfactual analysis, where we obtain corresponding GE effects on FDI.

We conclude the analysis of the impact of DTAs on trade by breaking our |$\mathrm{DEPTH}_{ij,t}$| variable into two count variables, |$\mathrm{WTO}$|-|$X_{ij,t}$| and |$\mathrm{WTO}+_{ij,t}$|⁠. The |$\mathrm{WTO}$|-|$X_{ij,t}$| commitments are commitments dealing with issues going beyond the current WTO mandate altogether, for example, commitments on labor standards, investments, or the environment. The |$\mathrm{WTO}+_{ij,t}$| commitments are commitments building on already agreed commitments at the multilateral level, for example, a further reduction in tariffs. The estimates for the two types of provisions, which appear in columns (6) and (7) of table 1, are very similar to each other. Thus, we do not find evidence for a differential impact of depth depending on whether the provisions are enforceable.

Our estimates of the effects of DTAs on FDI are presented in table 2. As discussed earlier, and similar to our trade specification, all estimates are obtained with the PPML estimator with three-way fixed effects, including origin-time, destination-time, and directional pair fixed effects. In addition, all specifications include indicator variables for WTO membership (⁠|$\mathrm{WTO}_{ij,t}$|⁠), bilateral investment treaties (⁠|$\mathrm{BIT}_{ij,t}$|⁠), and economic sanctions (⁠|$\mathrm{SANCT}_{ij,t}$|⁠). Similar to our approach with trade flows, to highlight the importance of DTAs and their provisions for FDI, we develop the estimation analysis sequentially.

The estimates in column (1) of table 2 include a single indicator variable, |$\mathrm{PTA}_{ij,t}$|⁠, that reflects the presence of a trade agreement of any type (e.g., deep or shallow) between i and j at time t. The main result from column (1) is that none of the effects of the policy variables in our model, including the impact of trade agreements and BITs, are statistically significant. A possible explanation for this result is that some of the significant determinants of FDI are country-specific variables on the origin and/or the destination side (e.g., corporate tax rate, corruption, bureaucratic red tape, and quality of institutions). However, such determinants are fully absorbed by the origin-time and destination-time fixed effects. Moreover, it is also possible that several time-invariant characteristics (e.g., bilateral distance and common official language) are important for FDI (cf. Eicher, Helfman, and Lenkoski 2012; Blonigen and Piger 2014). However, similar to the country-specific variables, these effects are also absorbed in our econometric model (by the pair fixed effects). Our finding on the insignificant impact of BITs may seem particularly strange; however, this result is common in the related literature, e.g., Lesher and Miroudot (2006) and Laget, Rocha, and Varela (2021).

Next, in column (2) of table 2 we allow for heterogeneous effects between shallow and deep trade agreements by using the same |$\mathrm{DTA}$| variable which we constructed for our trade regressions. Even though the estimate on |$\mathrm{DTA}_{ij,t}$| is positive, it is economically small and not statistically significant. Thus, unlike their significant impact on trade, our estimates suggest that DTAs per se do not promote FDI. One possible explanation for the insignificant estimates of the effects of PTAs and DTAs in our setting could be due to our inability to distinguish between horizontal and vertical FDI in our data. For example, it is plausible that PTAs could influence FDI due to increasing the incentive to undertake specifically vertical FDI, i.e., explore factor cost differences and disentangle (stages) of production and knowledge creation. If multinational firms rely heavily on global value chains (GVCs), one might expect that any agreement that lowers trade barriers would increase FDI by lowering the cost of shipping within GVCs. If horizontal FDI is seen as a substitute for exports, then one might expect PTAs to lower FDI, as they reduce the cost of trade and make it more appealing than FDI as a means of reaching consumers in the partner countries. Thus, it is indeed possible that the null effects of PTAs and DTAs on FDI are masking a positive effect for vertical FDI or GVCs that is offset by a negative effect for horizontal FDI.²¹

Crawford and Kotschwar (2020) and Laget, Rocha, and Varela (2021) provide evidence on the heterogeneous impact of different types of provisions on FDI. Motivated by their analysis and by the fact that some DTAs include provisions that are specifically targeted at investment, in our next specification (in column (3) of table 2), we isolate the effects of DTAs that include investment provisions. We expect that investment provisions may exert a direct positive effect on FDI. To construct the new covariate, we again rely on the World Bank’s DCRTA (cf. Hofmann, Osnago, and Ruta (2019) and Mattoo, Rocha, and Ruta (2020)), which includes 66 possible investment provisions. Based on this information, we construct a dummy variable, |$\mathrm{INV}_{ij,t}$|⁠, which takes a value of 1 if an agreement includes at least one investment provision, and it is equal to 0 otherwise. Thus, by construction, the observations that take a value of 1 in the |$\mathrm{INV}_{ij,t}$| indicator are a subset of the observations that are equal to 1 in the |$\mathrm{DTA}_{ij,t}$| dummy from column (2).

The main finding from column (3) is that we obtain a positive, sizable, and statistically significant estimate on |$\mathrm{INV}_{ij,t}$|⁠, which suggests that, on average, the PTAs with investment provisions in our sample have led to a 21.038 percent (std.err. 11.513) increase in FDI between their members. This result complements the findings from Laget, Rocha, and Varela (2021), who use firm-level data for the period 2003–2015 and obtain positive estimates of the effects of provisions related to “intellectual property rights” and “visa and asylum,” which vary between 32 and 50 percent, but do not find significant effects of investment provisions on FDI.²²

Given the insignificant estimates of all other policy variables in our specification, especially the BITs, and the general difficulty in the literature to identify policies that have significant effects on FDI (see Lesher and Miroudot (2006) and Laget, Rocha, and Varela (2021) for examples), we view our significant estimates of the impact of DTAs with investment provisions on FDI as an important result. Therefore, in the rest of the columns of table 2 we put this result to scrutiny by testing its robustness and exploring potential heterogeneity across several dimensions. We start by replacing the missing FDI values with zeros and by implementing the bias correction procedure of Weidner and Zylkin (2021), in columns (4) and (5), respectively. The new estimates on |$\mathrm{INV}_{ij,t}$| are very similar to our main result.

In column (6) of table 2 we use the DTA database to construct a continuous variable (⁠|$\mathrm{INV\_DEPTH}_{ij,t}$|⁠), which counts the number of investment provisions within each of the DTAs in our sample. The number of investment provisions across the DTAs in our sample varies between 7 and 41. The estimates from column (4) do not reveal a significant impact of the increase in the depth (number of provisions) on FDI. In fact, and pushing inference to the limit, our estimates suggest that the impact of additional provisions is negative. A possible interpretation of this result is that more investment provisions make the agreements more difficult to comply with. Even though our estimate on |$\mathrm{INV\_DEPTH}_{ij,t}$| is insignificant, we use it in combination with the positive estimates on |$\mathrm{INV}_{ij,t}$| to construct a continuous FDI response to the impact of DTAs. The effects are all positive and vary between 3.04 percent (std.err. 14.87) and 54.57 percent (std.err. 29.42), depending on the number of investment provisions.²³

Next we investigate the impact of DTAs with investment provisions over time. Specifically, in column (7) of table 2 we allow for heterogeneous effects in the 1990s vs. 2000s. Both estimates in column (7) are positive and sizable; however, only the estimate for the 2000s is statistically significant. A possible explanation for this result is that the investment provisions in the DTAs have become more effective over time. To obtain the estimates in fig. 3, we replace the single indicator for investment, |$\mathrm{INV}_{ij,t}$|⁠, with a series of leads and lags. The two main results from this figure are (a) that we do not find evidence for pre-trends; and (b) that it seems to take time for the effects of DTAs with investment provisions to materialize. While we find positive effects for all periods after entry into force of the DTA, the estimates only turn significant after 15 years. This may be the case because of the substantial uncertainty in our quantification reflected by the correspondingly large standard errors. It may also reflect the fact that, as compared to trade, investment relationships may take a longer time to shape.

We conclude the analysis of the effects of DTAs on FDI by investigating their impact on two specific countries—China and the United States.²⁴ Three results stand out from column (8) of table 2. First, we see that the estimate for the United States is large, positive, and statistically significant, suggesting that the DTAs with investment provisions have been particularly effective for this country. Second, the estimate for China is not statistically significant, suggesting that China has not benefited from its DTAs with investment provisions. Finally, the estimate for the remaining agreements with investment provisions, |$\mathrm{INV}_{ij,t}$|⁠, is positive and sizable, but it is no longer statistically significant. In combination with the results for China and the United States, this suggests that the average estimate on |$\mathrm{INV}_{ij,t}$| from column (3) is masking significant heterogeneity across countries and points to the benefits of obtaining country-specific effects.

In our last experiment, we allow for the country-specific effects of DTAs with investment provisions for China and the United States to vary depending on the direction of FDI. Three main results stand out from column (9) of table 2. First, we see that both estimates for the United States are positive and statistically significant. However, second, we notice an asymmetry in the impact of DTAs with investment provisions on the inward vs. outward FDI of the United States, in favor of the former. This is consistent with the observation that the United States is the largest receiver of FDI in our sample. Third, and most remarkable, the results are completely the opposite for China, where we obtain a very large, positive, and statistically significant estimate for this country’s outward FDI, but not statistically significant (and, in fact, negative) estimate for China’s inward FDI. We are not aware of existing estimates of asymmetric effects of DTAs on FDI depending on the direction of FDI flows and we believe that the stark differences that we obtain for China’s inward vs. outward FDI may have significant policy implications both for the negotiation and the implementation of DTAs with DFI provisions.

In sum, the analysis in this section demonstrated that while trade agreements do not necessarily promote trade and FDI on average, the impact of deep trade agreements on trade and the impact of deep trade agreements that include investment provisions on FDI is positive, statistically significant, and also heterogeneous over time and across countries too. We also offered evidence that deeper trade agreements (as measured by the number of provisions) lead to larger trade liberalization effects. However, we do not see evidence that increasing investment provisions in DTAs led to more FDI. Some of our results suggest that additional provisions and complexity may make DTAs less effective in promoting FDI. Next, we rely on the partial estimates from this section to obtain GE effects of DTAs on FDI.

4.2. Counterfactual Analysis

This section translates the partial equilibrium estimates from tables 1 and 2 into GE effects of DTAs on FDI. To this end, we rely on our structural trade and investment system . At the onset of this section, we point to a potential caveat with our GE analysis. Specifically, as noted earlier, the underlying theory is based on the assumption of non-rival technology FDI, while our data includes all/aggregate FDI flows. This gap, of course, has implications for the quantitative results. Therefore, the specific indexes we obtain and report in this section should be interpreted accordingly and with caution. Nevertheless, we believe that the main conclusions and policy implications we will draw in this section, e.g., about a disproportionately large impact on outward FDI, will remain qualitatively the same if applied to appropriate data on technology FDI.

We start the analysis by describing the steps we took to make system (1)–(8) operational for our purposes. For the counterfactual analysis, we use as a baseline the latest available year in our data set, which is 2011, determined by the availability of capital stock data. To perform the counterfactual analysis, we need to set values for the parameters. Some parameters are borrowed from the literature: (a) the elasticity of substitution is set equal to |$\sigma =6$|⁠,²⁵ (b) the consumer discount factor is set equal to |$\beta =0.98$| (Yao et al. 2012), and (c) the country-specific capital shares of production |$\alpha _j$| and the country-specific adjustment costs of capital |$\delta _{j}$| are calculated using the Penn World Tables and reported in columns (3) and (4) of table 3, respectively.

We calibrate other parameters to match the observed data. The share of technology capital of a country to all destinations as a share from total world technology capital (⁠|$\eta _{i}$|⁠) is calculated using |$FDI_{ij}^{\mathrm{value}}$|⁠:

The variable |$\phi _j$| is calculated using the relationship between inward FDI (⁠|$\mathrm{FDI}_j^{\mathrm{in}}=\sum _{i}\mathrm{FDI}_{ij}^{\mathrm{value}}$|⁠) and physical capital in the production function, along with FDI and physical capital data and data on the capital shares:

The exact values for |$\eta$| and |$\phi$| are given in columns (5) and (6) of table 3.

For the baseline, we calibrate bilateral trade frictions to the power of |$1-\sigma$|⁠, i.e., trade openness |$t_{ij}^{1-\sigma }$|⁠, using data on trade flows, income, and expenditure and solving equations (2) and (3) for given trade costs and calculating a new matrix |$t_{ij}^{1-\sigma }$| using equation (1) until convergence, where we normalize all internal trade costs and trade costs for one exporter to 1. Given trade costs, we can calculate the inward and outward multilateral resistance indexes using equations (2) and (3), respectively, where we set the inward MRT for Angola to 1.

The variable |$M_{j}$| is calibrated using data on income, FDI, and constructed MRTs and the following theory-consistent equation for technology capital:

With this, we can construct FDI openness (⁠|$\omega _{ij}$|⁠) using the following equation for FDI flows in values:²⁶

The variable |$A_{j}/\gamma _{j}$|⁠, the preference-adjusted technology, is calibrated using equations (4) and (5). As the values of domestic income and expenditure calculated from the trade data do not perfectly match up, we define |$\psi _j\equiv E_{j} / (Y_{j}+\eta _{j}\sum _{i \in \mathbb {N}_{ji,t}}\phi _iY_{i}- \phi _j Y_{j}\sum _{i \in \mathbb {N}_{ij,t}}\eta _{i})$| as an exogenous country-specific parameter that accounts for these trade imbalances. In the spirit of Dekle, Eaton, and Kortum (2007, 2008), we first eliminate all exogenous trade imbalances and take the equilibrium without trade imbalances as baseline.²⁷

To highlight the alternative channels through which DTAs affect FDI, and also to capitalize on the full set of our partial estimates, we perform two sets of experiments. First, we rely on our estimates of the dummy variables for DTAs and DTAs with investment provisions from column (3) of table 1 and column (3) of table 2, respectively. Then, we also obtain corresponding effects based on the estimates of the continuous depth variables from column (5) of table 1 and column (6) of table 2. We perform each of the two experiments in two steps. First, we change the vector of FDI frictions. Then, in addition, we change the vector of trade costs. As the DTAs are already in place, we perform an ex-post evaluation, i.e., we assume that in the baseline the agreement is in place and simulate the effect without DTAs as counterfactual. We then report the change from the baseline to the counterfactual, i.e., baseline value minus counterfactual value relative to the counterfactual value.

Given the main purpose of our analysis, and to keep the presentation of our results manageable, we focus the discussion of our counterfactual results on the percentage changes (between the baseline and the counterfactual scenarios) in the stocks of FDI per country. Specifically, we construct and report percentage changes in inward and outward FDI stocks, i.e., the percentage changes in technology capital used in total at home and technology capital from one country used abroad:

where the superscript b denotes baseline values with DTAs in place, and the superscript c the counterfactual situation without DTAs in place, and, consistent with our theory, inward FDI and outward FDI stocks per country can be calculated as follows:

Note that the inward FDI stock can be seen as the global technology stock applied locally, whereas the outward FDI stock is the usage of a country’s technology capital abroad. The variable |$\eta$| determines the usage of FDI abroad of one country, i.e., outward FDI stocks per country will change a lot if this share is large (i.e., |$\eta$| is large), even if the change in technology capital |$M_{i}$| is comparably small. This is due to the non-rival nature of FDI. This is not the case for inward FDI per country, where the |$\eta$|’s always sum to 1, and therefore changes in inward FDI are a weighted average of changes in the |$\omega _{ij}M_{i}$|⁠. In other words, the non-rival nature of FDI in our theory implies that any investment into technology FDI increases the stock of FDI available to all countries. The bilateral nature, i.e., that the outward FDI of one country is inward FDI for another country, does not hold up with this notion of FDI. Rather, as soon as an investment in non-rival FDI is undertaken, it can be used by every country in the world subject to friction. We believe that this knowledge flow is important and captured by the investment provisions.

To use a concrete example, consider a DTA with investment provisions signed between the United States and the European Union (EU). Because this DTA reduces the cost of FDI between the United States and EU countries, US firms invest more in technology capital, which cannot only be used in the United States, but also in all EU countries and other FDI destinations of the US firms. This leads to a greater increase in outward relative to inward FDI globally. Hence, an increase in technology FDI from the United States (i.e., an increase in the outward FDI of the United States), leads to an increase of inward FDI not only for one country but for all countries, even though to a different degree, depending on the FDI frictions (see equations (11) and (12)) and the share of technology capital of a country (as a share from total world technology capital). Whenever a country has a large share of FDI in the world, as is the case for China and the United States, any change in FDI frictions will have a comparably large effect on many FDI-receiving countries. The implication is that reducing frictions for countries that have large FDI outward flows, will lead to strong increases in inward FDI.²⁸

Our findings are reported in table 3, where the first column lists the ISO3 country codes for the countries in our sample, the second column the country names, and columns (7)–(10) report the results from the scenario based on the estimates of the dummy DTA variables.²⁹

The results in columns (7) and (8) of table 3 are obtained in response to a change in the bilateral FDI frictions between all countries that have signed a DTA with investment provisions that are based on our estimate of 0.191 (std.err. 0.095) from column (3) of table 2. There are several things noteworthy. First, both inward and outward FDI increase for most of the countries. For the countries that have signed a DTA, the effect for inward FDI is on average an about 2 percent increase, while it amounts to 42 percent for outward FDI. The large values for outward FDI are driven by the importance of China and the United States as the largest outward FDI countries. Their technology capital as a share of total world technology capital (i.e., their |$\eta$|’s) is about 18 percent (see table 3). Hence, their stocks are used substantially in many countries of the world (the exact usage at the bilateral level also depends on the FDI frictions |$\omega$|⁠). Even though China and the United States only increased their technology capital stock (M) by about 0.16 percent and 0.12 percent, respectively, the effect on their outward FDI stocks is large due to the huge share of their FDI in world FDI and the non-rival nature, allowing the technology capital to be used in all countries in the world simultaneously.

On the inward FDI side, we see the largest increases for Chile, Singapore, Peru, Thailand, the Philippines, Malaysia, Indonesia, Vietnam, the Dominican Republic, and the Republic of Korea. These are all countries that have many DTAs and also rely substantially on inward FDI. On the other end of the spectrum are countries that are hardly affected, either on the inward or on the export side, such as Sudan, Turkmenistan, the Islamic Republic of Iran, and Iraq. Those countries do not have many (or any) DTAs in place and are also relatively closed regarding FDI. Overall, we see a wide heterogeneity among countries. This is even more extreme for outward FDI, where the importance of the large outward FDI investors is very dominant.

The estimates in columns (9) and (10) of table 3 are obtained when, in addition to the change in bilateral FDI frictions, we also change the vector of bilateral trade frictions based on our estimate from column (2) of table 1. Relative to the scenario where only the bilateral FDI frictions are changed (i.e., the results presented in columns (7) and (8) of table 3), we see qualitatively a very similar picture and quantitatively an increase in both, inward and outward FDI. Specifically, on average, trade liberalization has contributed to a 0.8 percentage points (or about 46 percent) increase in inward FDI and about 11 percentage points (or about 26 percent) in outward FDI. These estimates reveal that trade liberalization via DTAs is an important channel to stimulate FDI, thus complementing the results from Anderson, Larch, and Yotov (2019), who show that FDI liberalization is important for trade.

table S4.1 in the supplementary online appendix reports estimates that are obtained based on the estimates of the continuous depth variables . Te estimates in columns (2) and (3) rely on the estimates from column (4) of table 2. Allowing for continuous depth leads to qualitatively and quantitatively similar results, with heterogeneous changes across countries. The estimates in columns (4) and (5) of table S4.1 are obtained when, in addition to the change in bilateral FDI frictions, we also change the vector of bilateral trade frictions based on our estimate from column (3) of table 1. Similar to the changes based on the estimates with the dummy variables, the additional allowance for changes in bilateral trade frictions leads to larger effects for inward and outward FDI.

Table S4.2 in the supplementary online appendix reports estimates obtained based on the estimates of the dummy variables but using a value of 4 instead of 6 for |$\sigma$|⁠. The estimates in columns (2) and (3) rely on the estimates from column (3) of table 2. Decreasing |$\sigma$| from 6 to 4 hardly changes the effects on inward and outward FDI if DTAs only lower FDI frictions. The estimates in columns (4) and (5) of table S4.2 are obtained when, in addition to the change in bilateral FDI frictions, we also change the vector of bilateral trade frictions based on our estimate from column (3) of table 1. In this case, we see an increase of about 0.5 percentage points of inward FDI and about 7 percentage points of outward FDI when |$\sigma$| is decreased from 6 to 4. A lower |$\sigma$| implies a lower substitutability of varieties. Hence, lowering trade costs leads to a larger change in trade flows when |$\sigma$| is lower. This reduces prices more and leads to a larger increase in inward and outward FDI.

Given the huge importance of China and the United States as investors in technology FDI and their corresponding large role in outward FDI, in table S4.3 in the supplementary online appendix we report estimates that are obtained based on the estimates of the dummy variables, but excluding China and the United States. All other assumptions are the same as for the results in table 3. In table S4.3 we report, after the country codes and country names, the values of the parameters |$\alpha$|⁠, |$\delta$|⁠, |$\eta$|⁠, and |$\phi$|⁠. The estimates in columns (7) and (8) rely on the estimates from column (3) of table 2. Excluding China and the United States from the data set leads to slightly larger average inward FDI effects for liberalizing countries and lower effects for non-liberalizing countries if DTAs only lower FDI frictions. Outward FDI is on average substantially smaller for the liberalizing countries, but more countries play a larger role now, like Japan, Korea, India, France, and Germany. This reflects the dominance of China and the United States for outward FDI. The estimates in columns (9) and (10) of table S4.3 are obtained when, in addition to the change in bilateral FDI frictions, we also change the vector of bilateral trade frictions based on our estimate from column (3) of table 1. As in the results before, lowering trade frictions alongside FDI frictions increases the effects on inward and outward FDI, highlighting again that trade and technology FDI are complementry, as discussed above.

To sum up, according to our analysis, the DTAs in force in 2011 contributed to about 3 percent of inward FDI in the world and about 50 percent of outward FDI. The latter is heavily driven by the fact that some countries have large stocks of FDI used in many countries in the world, multiplying the effect of any change in outward FDI of those countries due to changes in frictions.

5. Conclusion

The objective of this paper was to study the links between deep trade liberalization in the form of DTAs and FDI. To this end, we identified and decomposed three channels through which DTAs impact FDI. First, we obtained significant direct/partial equilibrium effects of DTAs and their investment provisions on FDI from a theory-motivated FDI gravity model. Second, we translated the partial estimates of the DTA effects on FDI into GE effects. This analysis highlighted the importance of the GE links between DTAs and FDI and uncovered significant asymmetries in the response of inward vs. outward FDI in our model. Finally, we performed a counterfactual analysis of the impact of deep trade liberalization on FDI, which revealed that, through their impact on trade, DTAs promote FDI additionally.

While, as discussed earlier, our counterfactual analysis is subject to criticism on the mismatch between the data used and the underlying theory, we believe that our conclusions about the disproportionately large impact on outward FDI would remain qualitatively the same if applied to appropriate data on technology FDI. We view this finding as novel and potentially important from a policy perspective, both for the negotiations of trade and investment agreements and for properly quantifying their implications. Moreover, we see significant potential in developing and utilizing data sets on global technology transfers that would generate more precise partial estimates and more informative GE analysis of the links between trade liberalization and FDI and lead to clearer policy recommendations. In addition to the theory on the intensive margin that we utilize here, we expect significant payoffs from developing theories that would capture the links between trade liberalization and the extensive margins (both domestic and international) of technology capital and its diffusion in the global economy.

Data Availability Statement

The data underlying this article and the codes to reproduce the results are available under the following link https://www.dropbox.com/scl/fo/rurwez2lzvqewqay9h3gs/AGO7T2k6gNzusg2XAFbsi-Q?rlkey=m84nzkmmb3od0nsgz3hcp3tcp&st=70vo8ab8&dl=0.

Author Biography

Mario Larch is a professor at the Department of Law, Business & Economics, University of Bayreuth, Universitätsstraße 30, 95447 Bayreuth, Germany; Centre d’Etudes Prospectives et d’Informations Internationales (CEPII), Paris, France; the ifo Institute and CESifo Research Network, Munich, Germany; and the Nottingham Centre for Research on Globalisation and Economic Policy (GEP), Nottingham, United Kingdom. His email address is [email protected]. Yoto Yotov is a professor at the School of Economics, LeBow College of Business, Drexel University and the ifo Institute and CESifo Research Network, Munich, Germany. His email address is [email protected]. We are grateful to Vanessa Alviarez for a very thoughtful and constructive discussion of our paper, and for her excellent suggestions during the World Bank’s “Deep Trade Agreements Conference: Effects beyond Trade.” We also thank Emily Blanchard, Keith Maskus, Gianluca Orefice, Nadia Rocha, and Michele Ruta for very useful comments and suggestions. This paper has benefited from support from the World Bank’s Umbrella Facility for Trade trust fund financed by the governments of the Netherlands, Norway, Sweden, Switzerland, and the United Kingdom. All errors are our own. A supplementary online appendix is available with this article at The World Bank Economic Review website.

Footnotes

References