Abstract
While much of the existing literature on foreign direct investment (FDI) emphasizes flows tied to real economic activity, a significant share—estimated at over 30% of the global FDI stock—is funneled through tax haven jurisdictions. Drawing on 2010 data from the International Monetary Fund (IMF) concerning FDI stock positions, this study investigates the geographic, historical, and institutional drivers that shape the emergence and distribution of offshore FDI. The findings reveal that offshore FDI, despite its lack of physical assets, responds to geographic distance similarly to real FDI. Additionally, these offshore flows are heavily concentrated along the lines of former colonial relationships, highlighting strong investment linkages between ex-imperial powers and their (former) colonies. Surprisingly, many OECD member countries—despite playing leading roles in global efforts to curb tax avoidance—harbor substantial levels of offshore FDI within their financial sectors. These trends point to the embedded and far-reaching nature of offshore FDI, illustrating its presence not only in advanced economies but also across low- and middle-income countries.
Keywords: Offshore FDI, Tax havens, Foreign direct investment, IMF FDI data, Colonial ties, OECD, Trade agreements, Distance sensitivity, Global capital flows, Tax evasion.
© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third-party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit https://creativecommons.org/licenses/by-nc-nd/4.0/
1. Introduction
According to a Special Report by The Economist on offshore finance, there are an estimated 50 to 60 tax havens globally. These jurisdictions collectively host over two million shell companies, alongside thousands of banks, investment funds, and insurance firms (Economist, 2013). The report estimates that more than 30% of global foreign direct investment (FDI) flows through these havens. Figure 1, which uses IMF data from the end of 2010, highlights all FDI stock positions exceeding US$50 billion. Among the 30 nations that serve as either origins or destinations for the world’s largest bilateral FDI flows are well-known tax havens such as Bermuda, the British Virgin Islands, and Cyprus, as well as Luxembourg, Ireland, and Austria. Remarkably, the outward FDI stock from the United Kingdom is nearly as high as that of the United States, with the Netherlands and Luxembourg close behind. Luxembourg’s outbound FDI even surpasses that of economic giants like Germany and France, while the British Virgin Islands records greater outward FDI than Australia, Brazil, Canada, and Mexico combined.
Understanding the global map of foreign direct investment (FDI) requires recognizing the pivotal role played by tax havens and offshore financial centers. These jurisdictions fundamentally challenge how we define and interpret FDI, especially in terms of its geographic distribution. According to UNCTAD (2007, 245), FDI is defined as “an investment involving a long-term relationship and reflecting a lasting interest and control by a resident entity in one economy (foreign direct investor or parent enterprise) in an enterprise resident in an economy other than that of the foreign direct investor (FDI enterprise or affiliate enterprise or foreign affiliate).” Yet, this definition becomes increasingly problematic in light of how easily shell entities can be established in tax havens. Can a paper company be said to reflect a “long-term relationship”? Can it truly be considered “resident” in the same way as a manufacturing plant?
This also calls into question the boundary between direct and portfolio investment. Conventionally, portfolio investment is considered passive and financially motivated, lacking the intent to exert control, whereas FDI implies strategic involvement. However, the use of tax havens to route FDI—often purely for fiscal efficiency—blurs this distinction significantly (Dicken, 2011). These arrangements also impede the ability to trace the true origins and destinations of capital. The opacity of offshore structures makes it extremely difficult to identify the ultimate sources and end points of FDI flows.
Surprisingly, economic geography has yet to fully engage with the role of tax havens in shaping FDI. While geographers have produced rich case studies on offshore financial centers, focusing on their political and institutional characteristics and roles in global capital flows (Roberts 1994, 1995; Cobb, 1998; Hudson, 2000; Warf, 2002), sustained interest waned in the 2000s. This occurred precisely when offshore finance became increasingly central to the functioning of the global economy (Sarre, 2007; Wainwright, 2011; Wójcik, 2012).
Meanwhile, research on the economic geography of FDI has certainly acknowledged the importance of taxation, institutions, and regulatory environments. However, it has seldom examined the role of offshore finance or tax havens, either empirically or conceptually (Dunning, 2001). The emergence of new datasets now allows researchers to empirically investigate the placement of tax havens within global FDI networks. This development offers a timely opportunity to integrate the previously separate literatures on FDI and offshore finance.
This paper seeks to do just that. Specifically, we aim to investigate the determinants of tax haven–related FDI stocks in non-tax haven countries (termed “offshore FDI”) in comparison with FDI stocks among non-haven countries (termed “real FDI”). We use newly available IMF data on year-end 2010 FDI stock positions. Inward FDI stocks located in tax havens themselves are excluded due to poor data availability. The analysis employs a Poisson Quasi-Maximum Likelihood (PQML) gravity model to regress both offshore and real FDI against five main groups of variables: host economy size and development level, geographic proximity, economic agreements, taxation features, and institutional indicators.
Due to data reliability issues—many tax havens do not have consistent or credible economic indicators—the focus of our analysis is limited to host-country characteristics and bilateral relationships. Home-country variables are excluded. This allows us to maintain consistent analytical standards and focus on measurable country-level and dyadic factors influencing FDI flows.
Previous research has explored the impact of taxation on FDI decisions, including the use of intermediary jurisdictions. However, few if any studies have attempted to systematically compare how such variables affect offshore versus real FDI. Our study helps fill this gap. For the purposes of this analysis, we use the terms “tax haven” and “offshore financial center” (OFC) interchangeably. We acknowledge that OFCs provide services beyond tax benefits, such as corporate secrecy, ease of registration, and light-touch regulation. However, in practice, these attributes are tightly bundled with tax advantages, making them difficult to isolate empirically. Thus, we focus instead on differentiating real and offshore FDI, rather than separating tax havens from broader offshore centers.
Our results yield several important insights. Offshore FDI is revealed to be a truly global phenomenon, not limited to high-income economies. Developing countries also both receive and serve as intermediaries in offshore FDI flows. While it might be assumed that intangible capital is less constrained by distance, offshore FDI is, in fact, just as responsive to geographic proximity as real FDI.
Historical colonial ties are a particularly strong determinant of offshore FDI, even though they have little influence on real FDI. This suggests that offshore financial connections are rooted in long-standing relationships between former imperial powers and their former colonies. Regarding international agreements, OECD membership reveals a dual character. Though the OECD champions global efforts to curb tax avoidance, its members are more likely to curtail offshore FDI flows from non-members than from within their own group—indicating possible internal leniencies.
Taxation-related variables are significant for both offshore and real FDI. Flows tend to follow paths that minimize withholding taxes, confirming that multinational firms actively structure transactions to maximize tax efficiency. We also find substantial evidence of third-country “treaty shopping,” where firms use intermediary countries to exploit favorable bilateral tax treaties.
Interestingly, a legacy of communism does not appear to significantly encourage offshore FDI. This finding challenge assumption that post-communist countries might be more attractive for offshore structuring due to weak regulatory frameworks.
Together, these results suggest that offshore and real FDI are shaped by both overlapping and divergent factors. They also underscore the importance of considering tax havens and offshore financial strategies in any comprehensive study of FDI patterns.
The remainder of the paper introduces the theoretical foundations for the determinants of offshore FDI in relation to real FDI, drawing on the relevant literature. The methodology section outlines the definitional criteria for tax havens, explains the structure of the PQML model, and describes the construction and sources of the key variables. Section 4 presents the main empirical findings, including robustness tests that explore alternative definitions and specifications. Section 5 summarizes the contributions of the study to our understanding of FDI, tax havens, and offshore finance.
2. Determinants of real and offshore FDI
We evaluate five clusters of explanatory variables as potential determinants of FDI: host economy size and development level, geographic proximity, economic agreements, taxation regimes, and institutional quality. Our variable selection draws on established empirical research (Blonigen 2005; Blonigen & Piger 2011) and aligns with Dunning’s eclectic paradigm (Dunning 1977, 1993), which proposes three primary motives behind FDI: market-seeking, efficiency-seeking, and resource-seeking. Market-seeking motivations are proxied by host GDP and GDP per capita; resource-seeking motives correspond to institutional variables; while proximity, economic cooperation, and tax factors primarily influence the cost structure of FDI.
Because offshore FDI is distinct from traditional flows, special attention is directed toward taxation and institutional factors. Tax havens commonly function as intermediaries in FDI routing, facilitating diversion and round-tripping—where investors send capital abroad and bring it back disguised as foreign investment. For instance, estimates suggest that between 15% and 50% of China’s reported inward FDI is simply round-tripped capital via the Cayman Islands and British Virgin Islands (Vlček 2010). Tax havens are employed both for rent extraction (via tax arbitrage and regulatory evasion) and for value creation (e.g., access to legal protections or specialized financial services) (Buckley et al. 2007). Moreover, global anti-tax avoidance initiatives led by the OECD and the EU underscore the relevance of economic agreement variables in shaping offshore FDI flows (Palan et al. 2010).
Below we introduce each cluster of variables and articulate hypotheses—grounded in prior research—about how they might differentially affect offshore versus real FDI.
1. Host economy size and development level
We include two control variables: total host-country GDP and host GDP per capita. Larger national economies typically offer more investment opportunities, so higher GDP should positively correlate with both offshore and real FDI. Prior studies (e.g., Blonigen 2005) consistently affirm the role of GDP within gravity-model frameworks.
The effect of GDP per capita is more nuanced. Real FDI, especially vertical, export-oriented investments, may be drawn to low-wage, low‑income countries; alternatively, horizontal, import-substituting real FDI might gravitate toward more developed markets with higher per-capita income. Empirical findings on this relationship remain ambiguous (Blonigen 2005). As for offshore FDI, while theory suggests that tax haven–related activity may be more prevalent in lower-income countries (Baker 2005; Epstein 2005), wealthier nations also possess more capital for round‑tripping and higher demand for offshore services. Thus, we include GDP per capita primarily as a control variable to isolate the effects of other determinants.
2. Geographic proximity, language, and colonial links
We consider three dimensions of ‘proximity’:
- Physical distance
- Shared language
- Colonial history
Virtually all empirical analyses confirm that larger geographic distance reduces real FDI (Blonigen 2005; Anderson 2011). Offshore FDI is trickier: since it involves paper-based or registration-driven flows rather than commercial operations, distance might be expected to matter less. However, financial geography scholars observe that even financial services remain spatially clustered—typically around key time zones (Roberts 1994; Desai et al. 2004; Clark & Wójcik 2007; Martin 1999). Moreover, there are well-known regional clusters of proximate economies served by nearby offshore centers—for example, France‑Luxembourg‑Monaco and Germany‑Luxembourg‑Switzerland‑Liechtenstein (Donaghy & Clark 2003; Sikka 2003). Hence, we hypothesize that distance will reduce real FDI more strongly than offshore FDI, yet it may still exert a meaningful friction on both.
A shared official language generally promotes real FDI (Blonigen & Piger 2011; Eicher et al. 2012), and may similarly facilitate offshore FDI via ease of contract law familiarity and financial intermediaries. Conversely, because many tax havens are English-speaking yet serve global clients, the effect may be diluted (Shaxson 2011).
Colonial affiliations—past or present—have been shown to influence offshore financial systems (Eden & Kudrle 2005; Clegg 2009; Palan et al. 2010), suggesting enhanced offshore flows between ex-colonial powers and their former dependencies. These ties likely also influence real FDI, though to a lesser extent.
3. Economic integration agreements: OECD, EU, WTO, and EIAs
We include four binary indicators:
- OECD membership
- EU membership
- WTO membership
- Economic Integration Agreements (EIAs; e.g., regional trade treaties including services provisions)
The OECD pioneered anti-tax avoidance initiatives like the 1998 “Harmful Tax Competition” agenda (OECD 1998) and coordinates anti–money laundering efforts via the FATF. If taken at face value, OECD membership should be negatively associated with offshore FDI. In contrast, the OECD fosters real investment flows, so an expected positive association exists for real FDI. Past critiques (Sanders 2002; Webb 2004; Eden & Kudrle 2005; Sharman 2006) suggest that OECD countries often internalize offshore activities, potentially resulting in higher offshore FDI within OECD membership.
The EU likewise pursues anti-avoidance policies, although its effectiveness is debated (Rawlings 2007; Hemmelgarn & Gaëtan 2009; Klautke & Wichernrieder 2010; Palan et al. 2010). EIAs—which often liberalize services—could facilitate offshore capital flows and raise offshore FDI (Roy et al. 2006). WTO membership nominally includes commitments to open financial services, but its actual impact on offshore FDI is likely minimal (Adlung 2009). Regarding real FDI, impacts are theoretically ambiguous: trade liberalization may encourage export-oriented vertical FDI or discourage import-substituting horizontal FDI (Helpman 1984; Markusen 1984; Blonigen 2005). Empirical outcomes are mixed (Blonigen 2002; Dunning 2002; Baltagi et al. 2008; Buthe & Milner 2008; Medvedev 2012).
4. Taxation and tax-related variables
Agglomerations of multinational activities are highly sensitive to tax policy. Traditional expectations are that higher host-country corporate income tax (CIT) rates will deter real FDI (Devereux et al. 2002; de Mooij & Ederveen 2003). For offshore FDI, higher domestic CIT may induce inflows to tax havens as investors seek to minimize tax liabilities.
Withholding taxes on dividends, royalties, and interest payments are also key determinants; reductions in these taxes should encourage both real and offshore FDI. Double Taxation Treaties (DTTs), which eliminate tax overlap, are expected to stimulate both types of FDI—but they may also contain information-exchange provisions that reduce incentives for offshore structuring (Blonigen & Davies 2004; Barthel et al. 2009). Prior empirical work on treaty shopping, third-country conduits, and intra-firm structuring (Desai et al. 2003; Weichenrieder & Mintz 2008; Weyzig 2012) reinforces the significance of these variables. To our knowledge, no prior global study has directly compared the coefficients of CIT, DTTs, and withholding rates between real and offshore FDI.
5. Institutional quality, communist legacy, and Controlled Foreign Corporation (CFC) rules
Strong rule of law is widely recognized as attracting real FDI (Wei 2000; Bénassy‑Quéré et al. 2007). Its relationship with offshore FDI is ambiguous: stronger institutions encourages real investment but can also make hosts less attractive for round-tripped or tax-motivated flows. Meanwhile, countries with weaker legal systems may see domestic capital routed offshore via havens.
We also test a dummy for communist legacy. Although institutional environments in transition economies may discourage real FDI, their histories of expropriation and underdeveloped legal systems may spur offshore transactions to escape domestic restrictions, facilitating “institutional arbitrage” (Buckley et al. 2007; Ledyaeva et al. 2013; Vlček 2013). This would imply a positive correlation between communist legacy and offshore FDI. In contrast, real inward FDI may be neutral or even negatively associated with such legacy, balanced by economic liberalization trends and the strategic importance of FDI-driven growth models.
Lastly, we include an indicator for Controlled Foreign Corporation (CFC) rules—laws designed to limit profit shifting to offshore subsidiaries. While theoretical expectations suggest CFC rules should reduce offshore FDI, empirical evidence suggests they are often weak or inconsistently enforced (Rawlings 2007; Palan et al. 2010).
Summary of hypotheses and rationale
Offshore FDI appears to be governed by a distinct logic from real FDI, rooted as much in tax and institutional arbitrage as in traditional investment motives. While several potential influences—like GDP, economic agreements, taxation, proximity, and institutions—are shared across both types, the expected signs and magnitudes of their effects diverge. Moreover, contradictory pressures present ambiguous hypotheses in some cases (e.g., income levels, OECD membership), highlighting the complexity of offshore flows. Nonetheless, these complexities reinforce the need for separate empirical treatment of offshore FDI rather than aggregating it with real FDI in economic analysis.
The remainder of the paper operationalizes these variables within a gravity framework using Poisson Quasi-Maximum Likelihood estimation on bilateral FDI stock data, distinguishes between real and offshore FDI flows based on a transparent classification of tax havens and offshore financial centers, and presents robustness checks under alternative model specifications. The paper closes with a discussion of how our findings contribute to the literature on FDI, tax havens, and offshore finance and what implications they hold for policy and future research.
3. Methodology and Data
Given the conceptual ambiguities surrounding the terms “tax haven” and “offshore financial center” (OFC), many scholars have resorted to what Senator Grassley once referred to as a “you know it when you see it” method of classification (US Senate, 2002). This study adopts an “expert consensus” approach, identifying tax havens based on their presence in a sufficient number of eleven authoritative lists compiled by different scholars, as aggregated by Palan et al. (2010; see Appendix A). To evaluate the robustness of our findings, we employ three thresholds of agreement—25%, 50%, and 75%—corresponding to jurisdictions appearing on at least 3, 6, and 9 of these lists, respectively. These cut-offs are referred to by their OFC “scores” (3, 6, and 9). Furthermore, we conduct two additional sensitivity checks: one excludes “large” tax havens—those with a GDP exceeding $200 billion USD (e.g., Hong Kong, Singapore, Switzerland, and Ireland)—from the score-6 group, while the other adds the Netherlands, recognized as a treaty shopping hub, to the same group.
To test hypothesized drivers of both offshore and real FDI, we apply a gravity model framework using Poisson Quasi-Maximum Likelihood (PQML) estimation. Gravity models posit that the intensity of bilateral relationships is: (1) directly proportional to the economic “mass” of the two entities (e.g., their GDP); (2) inversely proportional to the physical distance separating them; and (3) shaped by additional multiplicative factors (Anderson, 2011). While initially developed to model trade flows, gravity models have been widely validated in empirical studies across a range of international flows, including FDI (Blonigen, 2005). Ordinary Least Squares (OLS) estimation of log-linearized gravity models is problematic due to heteroskedasticity induced by Jensen’s inequality, which can lead to biased results (Santos Silva and Tenreyo, 2006). These authors propose PQML estimation as a solution, as it does not require log-transforming the dependent variable and permits the inclusion of zero-valued observations, which are common in FDI data. Accordingly, we use Stata 11’s Poisson regression routine with robust standard errors to estimate our models.
Our dependent variable is the bilateral FDI stock, sourced from the IMF’s Coordinated Direct Investment Survey (CDIS), which provides end-2010 data. This complements the earlier Coordinated Portfolio Investment Survey (CPIS), and, unlike the OECD dataset used in prior work, offers much broader country coverage. The 2010 CDIS reports inward FDI stocks from 83 host countries, covering investments originating from 245 distinct economies. This unprecedented breadth, which includes outward FDI stock data for many small OFCs, facilitates a more comprehensive empirical investigation into the determinants of offshore FDI.
Figure 2 presents inward offshore FDI stocks (using OFC score-6) as a share of GDP across the 83 host countries in the CDIS dataset. Numerous tax havens and treaty shopping hubs like the Netherlands report offshore FDI stocks exceeding 50% of GDP. Among non-haven economies, Russia and Chile display relatively high concentrations of offshore FDI, whereas regions like South Asia, Oceania, and the Americas exhibit generally lower levels.
Despite its global reach, the CDIS is constrained by several limitations. First, its time coverage is limited to a single year—2010—which hampers any effort to assess causal relationships, though our independent variables are unlikely to suffer from endogeneity with respect to FDI. Second, the inward FDI data relies on host-reported figures and excludes most tax havens, many of which either fail to report or provide questionable data. This underreporting precludes analysis of inward FDI into tax havens, prompting us to exclude tax haven economies from the host sample and focus exclusively on FDI into “real” economies.
We construct separate models for two dependent variables: “real FDI” (i.e., FDI flows between non-haven economies) and “offshore FDI” (i.e., FDI from tax havens into real economies). The independent variables fall into two categories: characteristics of the host country (e.g., GDP per capita) and bilateral features of the origin-host relationship (e.g., physical distance). To isolate the influence of host-related factors, we include fixed effects for origin economies, controlling for all origin-specific FDI determinants in line with Wei (2000). Our core PQML model specification estimates FDI from origin i to host j, with associated error term ϵij.
FDIᵢⱼ = β₀ + β₁(origin dummy)ⱼ + β₂ ln(GDP)ᵢ + β₃ ln([GDP/cap]ⱼ) + β₄ ln(distanceᵢⱼ) + β₅…(other variables) + ϵij
To validate the results, we also run models that incorporate host fixed effects, especially to assess the robustness of bilateral variables (Table 5). We analyze 15 key independent variables—expanded to 25 when including host-specific controls—grouped into five thematic areas: host economy size and development level, proximity, economic agreements, taxation, and institutional quality. Descriptive statistics for these variables are provided in Appendix B.
Host economic size and development are captured through total nominal GDP and GDP per capita, serving as proxies for market size and development status. These figures are drawn from the IMF World Economic Outlook Database (2012). Proximity variables include physical distance, common language, and colonial ties (both metropole-to-colony and colony-to-metropole). Distance measures are based on the great-circle calculation between financial centers (e.g., New York for the US). A shared language is recorded when at least 10% of both populations speak the same language, accounting for diaspora ties and historic connections such as those among post-Soviet states. Language data is sourced from CEPII (2012) and related archives. Colonial relationships are coded as dummy variables based on post-WWII ties.
Economic agreements are captured using four indicators: membership in the OECD, EU, WTO, and participation in Economic Integration Agreements (EIAs), including both bilateral and multilateral treaties such as NAFTA. EIA data is obtained from the WTO (2012). Taxation variables include the statutory corporate tax rate (from Ernst and Young, 2012), the presence of a double taxation treaty (DTT) as listed by UNCTAD (2012), and a zero withholding tax on dividends or interest paid to parent firms, based on data from Ernst and Young.
Lastly, institutional variables comprise the host country’s rule of law score (World Bank, 2012), a dummy for a history of communist governance, and another indicating the presence of Controlled Foreign Corporation (CFC) rules (Deloitte, 2012).
A distinctive feature of our approach is the use of “host control” variables for each bilateral factor. These represent host-level averages or memberships—for example, average bilateral distance or OECD membership. This extends the concept of “remoteness” from trade gravity models (e.g., Blonigen and Piger, 2011) and serves three purposes: it controls for omitted host-level influences, allows second-order effects like treaty shopping to be explored, and isolates the global influence of multilateral organizations beyond bilateral ties.
4. Results
Tables 1 through 5 present the outcomes of the primary models evaluated in this study. Tables 1 to 4 include models with origin fixed effects only, both with and without host control variables, across various OFC classification thresholds. In each of these tables, the first two columns report the results for real and offshore FDI determinants without host control variables, while the third and fourth columns include these controls. Tables 1, 2, and 3 correspond to OFC definition scores of 3, 6, and 9, respectively. Table 4 provides sensitivity analyses based on modifications to the score-6 sample—specifically, by excluding large OFCs and separately adding the Netherlands. Table 5 presents robustness checks that incorporate both host and origin fixed effects for OFC scores of 3, 6, and 9. The models demonstrate excellent fit, with pseudo R² values reaching as high as 0.94 in specifications that include both sets of fixed effects. A detailed interpretation of the results for each of the five categories of independent variables is provided in the subsequent discussion.
4.1. Host Economy Size and Level of Development
Host GDP consistently yielded a significant coefficient between 0.75 and 1.0 for both real and offshore FDI, suggesting that the “mass” effect associated with host economy size is similarly influential across both forms of investment. In contrast, host per capita GDP demonstrated divergent impacts on inward real versus offshore FDI. For real FDI, the relationship with host per capita GDP was generally negative, pointing to a pattern of investment driven by low labor costs or natural resource extraction in less developed economies. This aligns with conventional expectations about how differences in factor endowments influence capital returns. In the case of offshore FDI, a negative correlation with host per capita GDP was observed in models excluding host controls or fixed effects. However, this effect disappears once such controls or fixed effects are incorporated. This suggests that the lower levels of offshore FDI in developed economies are better explained by OECD membership than by income levels alone. Overall, we do not expect host per capita GDP to independently influence offshore FDI; any observed correlations likely reflect other institutional factors, such as the rule of law.
4.2. Proximity
Our models strongly confirm that real FDI is highly influenced by geographic proximity. Consistent with previous literature, many of our models (1.1.1, 2.1.1, 2.2.1, 5.2.1, 5.3.1) show that countries sharing a common language tend to engage in more real FDI. However, a key insight from our findings is that this linguistic effect weakens or disappears altogether in host control models. This suggests that the apparent importance of common language in promoting FDI may partly be a byproduct of countries with globally dominant languages, such as English or French, attracting more investment in general. When controls are introduced, the specific contribution of shared language becomes less clear, indicating that the relationship may be more associative than causative.
Colonial ties, too, reveal nuanced effects. Only the models incorporating host fixed effects (5.2.1 and 5.3.1) show that historical colonial relationships result in increased real FDI from colonizers to former colonies. There is no significant evidence of this influence operating in the reverse direction—from colonies to colonizers. While the direct impact of colonial history on real FDI may appear limited, this interpretation should be tempered by its strong indirect influence through language, which as noted, plays a more complex role than previously assumed.
In contrast, the dynamics of offshore FDI reflect an almost inverse relationship with language and colonial history. Common language does not show a consistent or significant influence on offshore FDI, likely due to the dominance of English in offshore financial services globally and among most OFCs. This linguistic ubiquity reduces variation across jurisdictions and dampens the measurable effect of shared language on offshore FDI flows.
However, colonial history plays a substantially larger role in shaping offshore FDI. Bilateral colonial ties are consistently and strongly associated with greater offshore FDI. This effect is bidirectional, even though few colonizers are classified as OFCs, with the notable exception of the UK in the score-3 models (1.1.2 and 1.2.2). These results lend robust quantitative support to earlier qualitative studies suggesting that many offshore financial centers maintain enduring connections to their former colonial powers—particularly the UK. This is especially evident in current or former British dependencies, where London remains a central financial hub. The persistent significance of colonial relationships in host control robustness models further underscores that these patterns are not merely attributable to the UK attracting FDI equally from all OFCs.
Indeed, the UK’s current and former territories collectively account for a substantial proportion of offshore FDI: 29% from the UK itself, 18% from current dependencies, and 22% from former colonies—totaling 69% of OFC outward FDI stock. This highlights the continued dominance of what Palan et al. (2010) refer to as “Britain’s second empire” in the global offshore financial architecture.
Although offshore FDI appears to be less responsive to physical distance than real FDI, the estimated coefficient ranges for both overlap, suggesting a comparable degree of distance sensitivity. This finding challenges the intuitive expectation that offshore FDI—being more intangible—would be less influenced by physical proximity. The likely explanation lies in the need for in-person travel by nominees, directors, and other offshore finance professionals, as well as client engagement. Surprisingly, time zone proximity does not significantly affect offshore FDI, even though it is commonly cited as a major factor in the regional clustering of offshore services.
Finally, our results align with Blonigen and Piger (2011) in showing a generally positive relationship between host remoteness—measured as average distance to all origin countries—and both types of FDI. In most models, the coefficients for host control variables (capturing remoteness) are approximately equal in absolute value to the bilateral distance coefficients. This symmetry supports the idea that FDI flows are shaped more by relative distance between countries than by absolute geographic distance. Consequently, our findings reinforce the importance of incorporating remoteness, host controls, or equivalent metrics into gravity model specifications for FDI.
4.3. Economic Agreements
Our model results reveal that initiatives tied to the OECD have had significant, albeit often unintended, consequences for offshore FDI, while comparable EU measures have had minimal impact. These outcomes lend support to arguments that the OECD, along with the Financial Action Task Force (FATF), has effectively created a stratified hierarchy among offshore financial centers (OFCs), particularly within OECD countries (Eden and Kudrle, 2005). Specifically, host OECD membership is associated with substantially lower offshore FDI inflows—up to 75% less—compared to mid- and lower-tier OFCs with definition scores between 3 and 8 (see Tables 1 and 2). However, this strong negative relationship vanishes when the sample is restricted to top-tier OFCs (definition score ≥ 9, Table 3), suggesting that leading centers have adapted more successfully to international regulatory pressures.
This divergence is in line with the “scissors effect” described by Sharman (2005), where smaller and less powerful OFCs have been disproportionately impacted by OECD and FATF pressures, while larger or better-connected OFCs have managed to avoid similar constraints. This pattern is further underscored by the composition of high-tier OFCs: 45% of those with a score of 9 or above are British or Dutch dependencies, compared to only 10% of OFCs in the 3–8 score range. Such findings reinforce the notion that these dependencies benefit from the political protection or patronage of their OECD metropoles, allowing them to weather the effects of regulatory scrutiny more effectively (Eden and Kudrle, 2005).
More importantly, our models reveal that shared OECD membership consistently correlates with significantly higher levels of offshore FDI—between 5 and 15 times more—from OECD-based OFCs compared to non-OECD OFCs. This outcome casts a critical light on the OECD’s anti-tax avoidance campaign, lending credibility to accusations of hypocrisy from both non-OECD officials and neutral observers. Importantly, this pattern is unique to offshore FDI; OECD-related variables had no significant influence on real FDI flows in our models (except in the case of definition score 9, which may overstate real FDI by misclassifying offshore flows).
Although some scholars, including Palan et al. (2010), have posited that EU anti-tax haven measures—particularly the EU Savings Directive—have been more effective than those of the OECD, our empirical findings suggest otherwise. While certain models using a score of 9 (3.1.2 and 3.2.2) and a score of 6 with major OFCs removed (model 4.1.2) show a negative relationship between shared EU membership and offshore FDI, these results are only weakly and inconsistently significant. The reduction appears limited to jurisdictions like Luxembourg, Cyprus, Malta, and Gibraltar—excluding Ireland and the UK—and may reflect the Directive’s emphasis on individual savings rather than corporate flows. The lack of effect on Ireland, a predominantly corporate OFC, aligns with this interpretation. Still, no EU-related variable exhibits an effect of comparable magnitude or robustness to that of OECD membership, supporting earlier conclusions that the EU Savings Directive has had limited measurable impact (Hemmelgarn and Gaeten, 2009; Klautke and Wichenrieder, 2010).
Regarding real FDI, the influence of EU membership is similarly ambiguous. While some host control models indicate that EU members attract less real FDI, these effects are weak and inconsistent. More generally, the impact of EU membership cannot be disentangled from that of broader economic integration agreements (EIAs), which include the EU. Our findings concerning EIAs reflect earlier counterintuitive observations: although bilateral FDI between trade agreement partners may decline in relative or absolute terms, countries with a higher number of EIAs tend to attract more FDI overall (Blomstrom and Kokko, 1997; Dunning, 2002; Levy Yeyati et al., 2002; Buthe and Milner, 2008). The most likely explanation is that trade liberalization reduces the need for market-seeking bilateral investment while simultaneously increasing the host’s attractiveness as an export platform for non-signatory countries.
As expected, EIAs appear to encourage bilateral offshore FDI as well, likely due to provisions related to financial services liberalization. However, the coefficients for bilateral and host EIA participation show instability in certain models. Specifically, when large OFCs such as Hong Kong, Singapore, Switzerland, and Ireland are excluded from the sample (model 4.1.2), the bilateral EIA effect disappears, while the host EIA coefficient turns significantly positive. This suggests that offshore FDI from smaller OFCs gravitates toward jurisdictions that have broader networks of EIAs—a pattern reminiscent of “EIA shopping,” analogous to treaty shopping in tax planning.
Of all economic integration variables analyzed, shared WTO membership shows the strongest and most consistent positive association with real FDI inflows. Nonetheless, host WTO membership alone displays a statistically significant negative effect on total inward real FDI, often outweighing the positive bilateral coefficient. This paradoxical finding echoes similar results reported by Neumayer (2007), suggesting that WTO participation, while fostering mutual investment among members, may be associated with broader structural or regulatory changes that dampen overall FDI inflows.
As for offshore FDI, shared WTO membership appears to reduce flows in models that do not account for host-level controls or fixed effects. However, once such controls are included, the bilateral WTO effect loses significance. Host WTO membership alone shows a mild but consistent negative association with offshore FDI from all sources. This relationship is, however, only marginally significant and somewhat difficult to interpret, especially given the parallel negative association found with real FDI. One possibility is that WTO participation involves broader commitments to financial transparency or liberalization that inadvertently discourage certain types of FDI. Regardless, our findings suggest that the WTO’s influence on both real and offshore FDI warrants closer scrutiny in future studies.
4.4. Taxation
The results for taxation-related variables strongly suggest that both real and offshore foreign direct investment (FDI) patterns are shaped by rational investor preferences aimed at maximizing tax efficiency. As anticipated, lower corporate income tax rates in host countries are associated with increased levels of real FDI inflows. This aligns with conventional expectations that lower tax burdens serve as a pull factor for investment into tangible economic activity. Conversely, no significant relationship was found between host corporate tax rates and offshore FDI inflows. One plausible explanation is that while high-tax jurisdictions might create stronger incentives for routing investment through OFCs, these same jurisdictions also tend to receive less total FDI overall, thereby neutralizing any net effect on offshore FDI volumes.
Additionally, variables measuring the presence of Double Taxation Treaties (DTTs) and zero bilateral withholding tax rates consistently display a positive correlation with both real and offshore FDI. This reinforces the view that multinational enterprises and high-net-worth individuals structure their cross-border operations with a keen eye on minimizing tax obligations—particularly concerning intra-firm capital transfers. More than just facilitating tax planning, DTTs may also contribute to enhanced legal clarity and predictability, making them attractive despite the potential for information exchange provisions to reduce privacy. This finding aligns with more recent literature that has observed a positive effect of DTTs on bilateral FDI, diverging from earlier studies that tended to report negative associations. Although some of these inconsistencies across the literature may result from methodological or data-related differences, they may also point to an evolving function of DTTs—from tools of regulatory constraint to instruments of financial normalization (Rawlings, 2007). The fact that our models at least partially control for withholding tax rates bolsters this interpretation; the remaining positive effect of DTTs likely reflects non-tax advantages such as legal protections or dispute resolution mechanisms.
Our host control tests further suggest that treaty and withholding tax shopping by third-party countries is a widespread practice. This is especially evident in the context of the withholding tax variable, where host-level controls tend to show greater significance and larger coefficients than their bilateral counterparts—both for real and offshore FDI. These findings provide some of the most robust empirical support to date for the phenomenon of treaty shopping, and they suggest that OFCs often rely on “enabling jurisdictions” when structuring flows to and from real economies. These enabling jurisdictions may not be tax havens themselves but possess favorable and extensive tax treaty networks that are leveraged for tax planning. Supporting this, our DTT host control tests reveal that while DTTs tend to promote bilateral offshore FDI when the OFC threshold is defined broadly (score ≥3), their effect shifts when focusing on top-tier OFCs (score ≥9). In such cases, the bilateral effect of DTTs becomes statistically insignificant, and the host-level effect turns significantly positive. This implies that the most prominent OFCs may derive their benefits from DTT networks primarily through intermediaries. Countries such as the Netherlands and the UK are familiar in this role, but our data also implicate lesser-known players like Hungary, Iceland, Sweden, and Norway—with Hungary and Iceland bordering on tax haven classification.
It is important to note, however, that our findings on taxation variables may be affected by endogeneity. Specifically, the tax structure in a jurisdiction might itself be influenced by the volume of existing FDI, complicating causal inference. Given the cross-sectional nature of our dataset, we cannot fully account for such dynamic interactions over time. As such, while our results offer compelling patterns, they should be interpreted as provisional and subject to further investigation using longitudinal data.
4.5. Institutions
The findings related to institutional variables were less definitive compared to other groups of variables analyzed. In line with much of the existing literature, stronger rule of law in host countries was consistently associated with increased levels of real FDI. This suggests that institutional stability and reliable legal frameworks remain key attractors for tangible investment. A few models (1.1.2 and 2.2.2) also detected a positive relationship between host rule of law and inward offshore FDI, which may imply a broader trend of one-way capital flight from high-risk, low-governance jurisdictions to more stable, developed economies. However, this association was only intermittently significant and should thus be interpreted with caution.
There was little robust evidence supporting the idea that a country’s history of communism has a meaningful influence on its ability to attract either real or offshore FDI beyond the effects already captured by its institutional quality, specifically rule of law. While the coefficient for host communist history tended to be negative for real FDI, its statistical significance was marginal at best (e.g., model 3.1.1). In the case of offshore FDI, the effect of host communist history was inconsistent in both direction and strength, further undermining claims that former communist countries are particularly prone to round-tripping behavior in FDI flows.
Evidence regarding Controlled Foreign Corporation (CFC) rules was similarly limited but somewhat suggestive. CFC rules—intended to prevent tax deferral through foreign subsidiaries—were negatively correlated with inward offshore FDI, but this effect reached significance in only one model (1.2.2). At the same time, CFC rules showed a generally positive relationship with real FDI, complicating interpretation. One possible explanation is that the CFC rules variable correlates with other unobserved factors associated with investor protections not entirely accounted for by the rule of law variable, thereby weakening confidence in the ability of the models to isolate CFC effects. Another possibility is that this same unobserved factor might be obscuring any suppressive effect of CFC rules on real FDI.
Importantly, because CFC rules affect only domestic investors, our methodology limits us to identifying their influence on round-tripping offshore FDI. While the institutional variables did not produce strongly negative results, the absence of consistently significant positive effects is telling. Despite theoretical emphasis on institutional “arbitrage” or “escape” as motivations for offshore investment, our analysis suggests that tax efficiency remains the dominant driver. Except for the OECD’s apparent “offshore club” effect, we find little systematic divergence in offshore FDI behavior between countries with stronger or weaker institutions.
5.Conclusions and implications
To the best of our knowledge, this is the first study to systematically compare the drivers of offshore and real foreign direct investment (FDI). In order to fulfill this objective, we utilized 2010 International Monetary Fund (IMF) data on bilateral FDI stocks, along with established classifications of tax havens and offshore financial centers (OFCs), to model offshore and real FDI stocks using five categories of explanatory variables: economic size and development, proximity, economic agreements, taxation, and institutional quality. The results reveal important geographical, historical, and political influences shaping offshore FDI, with significant implications for both academic literature and policy-making. From a geographical standpoint, offshore FDI displays a level of sensitivity to physical distance that mirrors that of real FDI. Contrary to expectations that time zone alignment might mitigate the importance of distance through easier communication, the findings indicate that time zone proximity has limited explanatory power. This suggests that offshore financial activity—while often perceived as placeless and virtual—may in fact depend heavily on face-to-face contact between offshore finance professionals and their clients. This human element of offshore transactions implies a physical rootedness in what appears to be a virtual phenomenon. Future qualitative research is needed to explore the interpersonal and institutional relationships that underpin offshore FDI. In terms of scope, our findings demonstrate that offshore FDI is a global phenomenon encompassing both developed and developing countries. We find no consistent link between a host country’s level of development—whether measured by GDP per capita or rule of law—and its ability to attract offshore FDI. This undermines the frequently made claim that offshore finance is uniquely associated with the developing world. It is worth noting, however, that our IMF dataset does not capture unilateral capital flight to tax havens and is likely biased toward more formal or legally sanctioned forms of FDI. As a result, the role of poorer countries in offshore financial flows may be underestimated. Historically, our analysis shows strong offshore FDI connections between former colonial powers and their overseas territories. This stands in contrast to communist legacies, which show little to no influence on offshore FDI flows. These results challenge the narrative that offshore financial activity is primarily driven by institutional weakness in transitional or developing economies. Instead, our findings align with interpretations of offshore finance as a form of neo-imperial economic structure—what some have termed “Britain’s second empire”—where OFCs are effectively managed from onshore financial hubs, especially in the former colonial metropoles. The City of London, for instance, continues to play a dominant role in coordinating the financial activities of UK-linked OFCs (Palan et al., 2010). In essence, offshore finance is not located on the margins of the global economy but is embedded at its very core. As Maurer (2008) emphasizes, “far from being a marginal or exotic backwater of the global economy, offshore in many ways is the global economy” (p. 160, original emphasis). From a political perspective, our most significant finding concerns the unintended consequences of OECD-led anti-tax haven campaigns. While these initiatives were ostensibly designed to curtail offshore tax avoidance, they appear instead to have reshaped the offshore financial system into a hierarchical structure. OECD-member OFCs—or “inside renegades”—have largely benefited from these reforms, while smaller, non-OECD OFCs have suffered significant losses in market share (Eden and Kudrle, 2005). This trend confirms long-standing criticisms of OECD hypocrisy, voiced both by non-OECD jurisdictions and neutral analysts. In contrast, EU initiatives, including measures like the EU Savings Directive, appear to have had negligible effects on offshore finance as of the 2010 data. These findings carry important policy implications. At a detailed level, our result that the tax advantages provided by Double Taxation Treaties (DTTs) outweigh any deterrents created by enhanced scrutiny over tax avoidance strategies lends further support to the view that DTTs have transitioned from tools of offshore suppression to mechanisms of normalization and legitimization (Rawlings, 2007). Simultaneously, although governments have aimed to curb third-country treaty shopping, we find evidence that such practices remain widespread for both real and offshore FDI flows. More broadly, the truly global reach of offshore FDI underscores the need for coordinated international policy responses. It challenges simplistic narratives in which wealthy nations are cast as reformers correcting financial misconduct localized in the developing world. Instead, our results suggest that the architecture of offshore finance is sustained and, to some extent, orchestrated by the financial systems of developed countries—particularly the United Kingdom and its former colonies. Therefore, the responsibility for leading reform efforts likely rests with these same actors, especially the UK. This paper contributes to the economic geography literature in two important ways. First, by providing a systematic, large-sample quantitative analysis, it complements the predominantly case study-driven research on offshore finance. Second, by illustrating the extensive reach and internal logic of offshore FDI—one that is distinct yet thoroughly embedded in spatial and historical structures—it enhances our understanding of the economic geography of FDI more broadly. FDI remains a foundational element of globalization, whether conceptualized through global value chains, production networks, or other frameworks. Further research into offshore FDI is essential for deepening our understanding of the intersection between globalization and financialization—arguably two of the defining trends of the past four decades.
6. References
Adlung, R. (2009) Service liberalization from a WTO/GATS perspective: in search of volunteers. WTO Staff Working Paper ERSD-2009-05. Available at: https://www.wto.org/english/res_e/reser_e/ersd200905_e.pdf
Anderson, J. E. (2011) The gravity model. NBER Working Paper No. 16576. Available at: https://www2.bc.edu/~anderson/GravityModel.pdf
Anderson, J. E. and van Wincoop, E. (2001) Gravity with gravitas: a solution to the border puzzle. NBER Working Paper No. 8079. Available at: http://www.nber.org/papers/w8079.pdf?new_window=1
Baker, R. W. (2005) Capitalism’s Achilles Heel: Dirty Money and How to Renew the Free-Market System. London: John Wiley and Sons.
Baltagi, B. H.,Egger, P., Pfaffermayr, M. (2008) Estimating regional trade agreement effects on FDI in an interdependent world. Journal of Econometrics, 145: 194-208.
Barthel, F., Busse, M., Neumayer, E. (2009) The impact of double taxation treaties on foreign direct investment: evidence from large dyadic panel data. Contemporary Economic Policy, 28: 366-377.
Bénassy-Quéré, A., Coupet, M., Mayer, T. (2007) Institutional determinants of foreign direct investment. The World Economy, 30: 764-782.
Blomstrom, M. and Kokko, A. (1997) Regional integration and foreign direct investment. NBER Working Paper 6019. Available at: http://www.nber.org/papers/w6019.pdf?new_window=1
Blonigen, B. (2002) Tariff-jumping antidumping duties. Journal of International Economics, 57: 31-50.
Blonigen, B. (2005) A review of the empirical literature on FDI determinants. Atlantic Economic Journal, 33: 383-403.
Blonigen, B. A. and Davies, R. B. (2004) The effects of bilateral tax treaties on U.S. FDI activity. International Tax and Public Finance, 11: 601-622.
Blonigen, B. A. and Piger, J. (2011) Determinants of foreign direct investment. NBER Working Paper No. 16704. Available at: http://www.nber.org/papers/w16704
Buckley, P. J., Clegg, L. J., Cross, A. R., Liu, X., Voss, H., Zheng, P. (2007) The determinants of Chinese outward foreign direct investment. Journal of International Business Studies, 38: 499-518.
Buthe, T. and Milner, H. V. (2008) The politics of foreign direct investment into developing countries: increasing FDI through international trade agreements? American Journal of Political Science, 52: 741– 762.
Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) (2012) GeoDist [database]. Retrieved 15 October from: http://www.cepii.com/distance/geo_cepii.xls
Clark, G. L. and Wójcik, D. (2007) The Geography of Finance: Corporate Governance in the Global Marketplace. Oxford: Oxford University Press.
Clegg, P. (2009) Governing the UK Caribbean overseas territories: A Two-Way Perspective. In P. Clegg and E. Pantojas-Garcia (eds) Governance in the Non-independent Caribbean: Challenges and Opportunities in the 21st Century. Kingston: Ian Randle.
Cobb, S. C. (1998) Global finance and the growth of offshore financial centers: the Manx experience. Geoforum, 29: 7-21.
de Mooji, R. A. and Ederveen, S. (2003) Taxation and foreign direct investment: A synthesis of empirical research. International Tax and Public Finance, 10: 673-93.
Deloitte (2012) Taxation and Investment Guides and Country Highlights [database]. Available at: http://www.deloitte.com/taxguides [Accessed 4 November 2012].
Desai, M. A., Foley, C. F., Hines, J. R. (2003) Chains of ownership, regional tax competition, and foreign direct investment. In H. Herrmann, and R. Lipsey, R. (eds) Foreign Direct Investment in the Real and Financial Sector of Industrial Countries, pp. 61-98. Heidelberg: Springer Verlag
Desai, M. A., Foley, C. F., Hines, J. R. (2004) Economic effects of regional tax havens. NBER Working Paper No. 10806. Available at: http://www.nber.org/papers/w10806.pdf?new_window=1
Devereux, M. P., Griffith, R., Klemm, A. (2002) Corporate income tax reforms and international tax competition. Economic Policy, 17: 449-495.
Dicken, P. (2011) Global shift: Mapping the Changing Contours of the World Economy (6th edn.). Sage: London.
Donaghy, M. and Clarke, M. (2003) How far offshore? Monaco as a financial centre. Competition and Change, 7: 3-21.
Dunning, J. (1977) Trade, location of economic activity and the MNE: a search for an eclectic approach. In B. Ohlin, P. O. Hasselborn, P. M. Wijkmon (eds) The International Location of Economic Activity, pp. 395-418. London: Macmillan.
Dunning, J. (1993) Multinational Enterprises and the Global Economy. London: Sage. Eden, L. and Kudrle, R. T. (2005) Tax havens: renegade states in the international tax regime? Law & Policy, 27: 100–127.
Dunning, J. (2001) The eclectic (OLI) paradigm of international production: past, present, future. International Journal of the Economics of Business, 8: 173-190.
Dunning, J. H. (2002) The impact of the completion of the European internal market on FDI. In J. H. Dunning (ed) Regions, Globalization and the Knowledge-Based Economy. Oxford: Oxford University Press.
Eicher, T. S., Helfman, L., Lenkoski, A. (2012) Robust FDI determinants: Bayesian model averaging in the presence of selection bias. Journal of Macroeconomics, 34: 637-651.
Epstein, G. A. (2005) Capital Flight and Capital Controls in Developing Countries. Cheltenham: Edward Elgar Publishing.
Ernst and Young (2012) 2012 Worldwide Corporate Tax Guide [database]. Available at: http://www.ey.com/GL/en/Services/Tax/Worldwide-Corporate-Tax-Guide—Country-list [Accessed 1 November 2012].
Helpman, E. B. (1984) A simple theory of international trade with multinational corporations. Journal of Political Economy, 92: 451-71.
Hemmelgarn, T. and Nicodeme, G. (2009) Tax co-ordination in Europe: assessing the first years of the EUSavings Taxation Directive. EC Directorate-General for Taxation and Customs Union Working Paper No. 18. Available at: http://ec.europa.eu/taxation_customs/resources/documents/taxation/geninfo/economicanalysis/taxpapers/taxationpaper18.pdf
Hudson, A. (2000) Offshoreness, globalization and sovereignty: a postmodern geo-political economy? Transactions of the Institute of British Geographers, 25: 269-283.
IMF (2012) World Economic Outlook Database (April 2012 Edition). Available at: http://www.imf.org/external/pubs/ft/weo/2012/01/weodata/index.aspx [Accessed 15 October 2012].
Klautke, T. and Weichenieder, A. J. (2010) Interest income tax evasion, the EU Savings Directive and capital market effects. Fiscal Studies, 31: 151-170.
Ledyaeva, S., Karhunen, P., Whalley, J. (2013) If foreign investment is not foreign: round-trip vs. genuine foreign investment in Russia. Working paper. Available at: https://tippie.uiowa.edu/economics/tow/papers/ledyaeva-spring2013.pdf
Levy Yeyati, E., Stein, E., Daude, C. (2002) Regional integration and the location of FDI. Inter-American Development Bank Working Paper. Available at: http://www.iadb.org/intal/intalcdi/PE/2009/03494a02.pdf
Markusen, J. R. B. (1984) Multinationals, multi-plant economies, and the gains from trade. Journal of International Economics, 16: 205-26.
Martin, R. (ed) (1999) Money and the Space Economy. Chichester: John Wiley & Sons.
Maurer, B. (2008) Re-regulating offshore finance? Geography Compass, 2: 155-175.
Medvedev, D. (2012) Beyond trade: the impact of preferential trade agreements on FDI inflows. World Development, 40: 49-61.
Neumayer, E. (2007) Do double taxation treaties increase foreign direct investment to developing countries? Journal of Development Studies, 43: 1501-1519. OECD (1998) Harmful Tax Competition: An Emerging Global Issue. Paris: OECD. Available at: http://www.oecd.org/tax/harmfultaxpractices/1904176.pdf
Palan, R., Murphy, R., Chavgneux, C. (2010) Tax Havens: How Globalization Really Works. Ithaca: Cornell University Press.
Rawlings, G. (2007) Taxes and transnational treaties: responsive regulation and the reassertion of offshore sovereignty. Law & Policy, 29: 51-66.
Roberts, S. (1994) Fictitious capital, fictitious spaces: the geography of offshore financial flows. In R. Martin, N. Thrift, S. Corbridge (eds) Money, Power and Space. Oxford: Blackwell.
Roberts, S. (1995) Small place, big money: The Cayman Islands and the international financial system. Economic Geography 71: 237-256.
Roy, M., Marchetti, J., and Lim, H. (2006) Services liberalization in the new generation of preferential trade agreements (PTAs): how much further than the GATS? WTO Staff Working Paper ERSD-2006-07. Available at: https://www.wto.org/english/res_e/reser_e/ersd200607_e.pdf
Sanders, R. (2002) The fight against fiscal colonialism: the OECD and small jurisdictions. The Round Table, 91: 325-348.
Santos Silva, J. M. C. and Tenreyro, S. (2006) The log of gravity. The Review of Economics and Statistics, 88: 641-658.
Sarre, P. (2007) Understanding the geography of international finance. Geography Compass, 1: 1076- 1096.
Sharman, J. C. (2005) South Pacific tax havens: from leaders in the race to the bottom to laggards in the race to the top? Accounting Forum, 29: 311–323.
Sharman, J. C. (2006) Havens in a Storm: The Struggle for Global Tax Regulation. Ithaca: Cornell University Press.
Shaxson, N. (2011) Treasure Islands: Tax Havens and the Men who Stole the World. London: Bodley Head.
Sikka, P. (2003) The role of offshore financial centres in globalization. Accounting Forum, 27: 365-399. The Economist (2013). Special report: Offshore finance. 16 February.
UNCTAD (2007) World Investment Report 2007. Available at: unctad.org/en/Docs/wir2007_en.pdf
UNCTAD (2012) Country-Specific Lists of Double Taxation Treaties [database]. Available at: http://unctad.org/en/Pages/DIAE/International%20Investment%20Agreements%20(IIA)/Countryspecific-Lists-of-DTTs.aspx [Accessed 1 November 2012].
Vlcek, W. (2010) Byways and highways of direct investment: China and the offshore world. Journal of Current Chinese Affairs, 39: 111-142.
Vlcek, W. (2013) From Road Town to Shanghai: situating the Caribbean in global capital flows to China. The British Journal of Politics & International Relations (forthcoming).
Wainwright, T. (2011) Tax doesn’t have to be taxing: London’s ‘onshore’ finance industry and the fiscal spaces of a global crisis. Environment and Planning A, 43: 1287-1304.
Warf, B. (2002) Tailored for Panama: offshore banking at the crossroads of the Americas. Geografiska Annaler, 84: 33-47.
Webb, M. (2004) Defining the boundaries of legitimate state practice: norms, transnational actors and the OECD’s project on harmful tax competition. Review of International Political Economy, 11: 787-827.
Wei, S. (2000) How taxing is corruption on international investors? The Review of Economics and Statistics, 82: 1-11.
Weichenrieder, A. J. and Mintz, J. (2008) What determines the use of holding companies and ownership chains?” Oxford University Centre for Business Taxation Working Paper Series No. 803. Available at: http://www.sbs.ox.ac.uk/centres/tax/symposia/Documents/2007/A%20Weichenrieder%20Holding _companies_15March07.pdf
Weyzig, F. (2012) Tax treaty shopping: structural determinants of foreign direct investment routed through the Netherlands. International Tax and Public Finance, September.
Wójcik, D. (2012) Where governance fails: advanced business services and the offshore world. Progress in Human Geography, doi: 10.1177/0309132512460904
World Bank (2012) World Governance Indicators [database]. Available at: http://info.worldbank.org/governance/wgi/index.asp [Accessed 15 October 2012].
WTO (2012) List of all RTAs [database]. Available at: http://rtais.wto.org/UI/PublicAllRTAList.aspx [Accssed 9 October 2012].
Table 1: OFC Definition Score >= 3
| Variable | 1.1.1: Real | 1.1.2: OFC | 1.2.1: Real | 1.2.2: OFC |
| Pseudo R² | 0.88 | 0.83 | 0.90 | 0.90 |
| Host Economy Size | ||||
| ln(GDP) (host) | 0.81**** | 0.93**** | 0.82**** | 0.71**** |
| Level of Development | ||||
| ln(GDP/cap) (host) | (0.37) *** | (0.97) *** | (0.34) *** | -0.13 |
| Proximity | ||||
| ln(Distance) | (0.64) **** | (0.58) *** | (0.79) **** | (0.33) ** |
| ln(Distance) (host) | 1.30**** | 1.28* | ||
| Common Language | 0.57*** | -0.25 | 0.29 | 0.20 |
| Common Language (host) | 1.92**** | -0.48 | ||
| Colonial Links | ||||
| Metro–Colony | 0.19 | 2.28*** | 0.18 | 2.34**** |
| Colony–Metro | -0.89 | 2.32**** | -0.64 | 1.72**** |
| Colony–Metro (host) | (10.32) **** | 42.71**** | ||
| Economic Agreements | ||||
| EIAs | 0.24 | 0.61 | -0.38 | 0.73** |
| EIAs (host) | 8.65**** | 1.17 | ||
| EU | (0.67) *** | -0.87 | -0.28 | -0.45 |
| EU (host) | (0.96) * | -0.37 | ||
| OECD | 0.046 | 1.79**** | 0.28 | 2.69**** |
| OECD (host) | 0.31 | (1.12) ** | ||
| WTO | 0.25 | (1.84) *** | 1.43*** | -0.12 |
| WTO (host) | (2.91) **** | (1.93) ** | ||
| Taxation | ||||
| ln (Tax Rate) (host) | (1.50) **** | (1.56) ** | (1.30) **** | -0.098 |
| DTTs | 0.32 | 0.84* | 0.56** | 0.78** |
| DTTs (host) | 1.68 | 1.78* | ||
| No Withholding Tax | 0.73**** | 1.17**** | 0.43** | 0.73** |
| No Withholding (host) | 1.35**** | 1.11**** | ||
| Institutions | ||||
| Rule of Law (host) | 0.66*** | 0.99** | 0.24** | 0.061 |
| Communism (host) | -0.21 | 0.18 | -0.31 | (0.54) ** |
| CFC Rules (host) | 0.38** | 0.46 | 0.14 | (0.77) * |
| Constant | 14.79**** | 14.57*** | 4.79 | -6.48 |
Table 2: OFC Definition Score >= 6
| Variable | 2.1.1: Real | 2.1.2: OFC | 2.2.1: Real | 2.2.2: OFC |
| Pseudo R² | 0.89 | 0.81 | 0.92 | 0.85 |
| Host Economy Size | ||||
| ln(GDP) (host) | 0.76**** | 0.90**** | 0.84**** | 0.96**** |
| Level of Development | ||||
| ln(GDP/cap) (host) | (0.30) *** | (0.60) ** | (0.51) **** | -0.43 |
| Proximity | ||||
| ln(Distance) | (0.63)**** | (0.76)**** | (0.79)**** | (0.68)**** |
| ln(Distance) (host) | 1.00*** | 0.56 | ||
| Common Language | 0.70*** | (0.85)** | 0.60**** | (1.04)**** |
| Common Language (host) | 2.16**** | 0.23 | ||
| Colonial Links | ||||
| Metro–Colony | 0.47 | (omitted) | 0.51 | (omitted) |
| Colony–Metro | -0.17 | 0.62 | 0.36 | 1.92*** |
| Colony–Metro (host) | (8.68)**** | (3.29)**** | ||
| Economic Agreements | ||||
| EIAs | 0.17 | 0.86** | (0.48)* | 0.71* |
| EIAs (host) | 9.17**** | 2.00 | ||
| EU | (0.65)*** | (0.95)* | -0.28 | -0.45 |
| EU (host) | (1.04)** | -1.06 | ||
| OECD | -0.041 | 2.04**** | 0.44* | 2.63**** |
| OECD (host) | 0.23 | (1.56)*** | ||
| WTO | 0.30 | (1.35)*** | 1.08** | -0.32 |
| WTO (host) | (2.47)**** | (1.72)* | ||
| Taxation | ||||
| ln(Tax Rate) (host) | (1.20)**** | -0.23 | (1.69)**** | 0.31 |
| DTTs | 0.31 | 0.83 | 0.62** | 1.10* |
| DTTs (host) | 1.50 | -0.13 | ||
| No Withholding Tax | 0.73**** | 0.94**** | 0.43** | 0.32 |
| No Withholding (host) | 1.37**** | 1.30**** | ||
| Institutions | ||||
| Rule of Law (host) | 0.57**** | 0.22 | 0.37*** | 0.82** |
| Communism (host) | -0.17 | 0.58 | (0.34) * | 0.55 |
| CFC Rules (host) | 0.50*** | 0.19 | 0.42*** | -0.14 |
| Constant | 13.41**** | 8.73** | 9.43*** | 1.78 |
Table 3: OFC Definition Score >= 9
| Variable | 4.1.1: Real | 4.1.2: OFC | 4.2.1: Real | 4.2.2: OFC |
| Pseudo R² | 0.85 | 0.77 | 0.89 | 0.82 |
| Host Economy Size | ||||
| ln (GDP) (host) | 0.58**** | 0.95**** | 0.75**** | 0.96**** |
| Level of Development | ||||
| ln(GDP/cap) (host) | 0.26 | (0.47)* | 0.28 | -0.36 |
| Proximity | ||||
| ln(Distance) | (0.60) **** | (0.57)*** | (0.77)**** | (0.53)*** |
| ln(Distance) (host) | 0.58 | 0.65 | ||
| Common Language | 0.19 | -0.48 | 0.12 | -0.5 |
| Language (host) | 1.13*** | 0.54 | ||
| Colonial Links | ||||
| Metro–Colony | 0.034 | (omitted) | 0.091 | (omitted) |
| Colony–Metro | -0.29 | 0.58 | 0.24 | 1.77** |
| Colony–Metro (host) | (4.40)*** | (5.27)**** | ||
| Economic Agreements | ||||
| EIAs | 0.058 | 0.81** | -0.32 | 0.53 |
| EIAs (host) | 0.24 | 1.56 | ||
| EU | (0.44)* | (1.58) ** | -0.27* |
Table 4: OFC Definition Score >= 6; No Large OFCs^ & Netherlands as OFC
| Variable | 5.1.1: No Host Controls | 5.1.2: Host Controls | 5.2.1: No Host Controls | 5.2.2: Host Controls |
| Dependent Variable | OFC | OFC | OFC | OFC |
| Pseudo R² | 0.80 | 0.84 | 0.87 | 0.89 |
| Host Economy Size | ||||
| ln(GDP) (host) | 0.77**** | 0.95**** | 0.94**** | 0.86**** |
| Level of Development | ||||
| ln(GDP/cap) (host) | -0.46 | -0.56 | -0.27 | 0.095 |
| Proximity | ||||
| ln(Distance) | (0.86)*** | (1.04)**** | (0.37)**** | (0.43)**** |
| ln(Distance) (host) | 1.55* | 0.54 | ||
| Common Language | (1.20)* | (1.63)**** | 0.29 | 0.065 |
| Common Language (host) | -0.53 | 0.57 | ||
| Colonial Links | ||||
| Metro–Colony | (omitted) | (omitted) | 0.62* | 1.23**** |
| Colony–Metro | 0.34 | 2.54*** | 1.31** | 1.47*** |
| Colony–Metro (host) | (3.03) *** | (1.29)** | ||
| Economic Agreements | ||||
| EIAs | 0.29 | 0.0034 | 0.52* | 0.44 |
| EIAs (host) | 4.87* | 1.56 | ||
| EU | -0.74 | (1.15)** | -0.28 | 0.24 |
| EU (host) | -1.68 | -1.13 | ||
| OECD | 2.65**** | 3.29**** | 1.50**** | 2.16**** |
| OECD (host) | (1.21)** | (1.53)**** | ||
| WTO | (1.51)** | -0.34 | -0.54 | -0.11 |
| WTO (host) | (2.17)** | -0.54 | ||
| Taxation | ||||
| ln(Tax Rate) (host) | 0.18 | 0.98 | -0.39 | -0.15 |
| DTTs | 1.05 | 1.38* | 0.60** | 0.84** |
| DTTs (host) | 0.25 | 0.057 | ||
| No Withholding Tax | 1.30**** | 0.82** | 0.69*** | 0.48** |
| No Withholding (host) | 1.02*** | 0.45 | ||
| Institutions | ||||
| Rule of Law (host) | 0.086 | 0.61 | -0.29 | -0.13 |
| Communism (host) | 0.42 | 0.36 | 0.49 | 0.43 |
| CFC Rules (host) | -0.084 | -0.42 | -0.078 | -0.14 |
| Constant | 8.22 | -4.3 | 3.48 | -3.23 |
Table 5: Host Fixed Effects Tests
| Variable | 6.1.1: Real (≥3) | 6.1.2: OFC (≥3) | 6.2.1: Real (≥6) | 6.2.2: OFC (≥6) | 6.3.1: Real (≥9) | 6.3.2: OFC (≥9) |
| Pseudo R² | 0.93 | 0.93 | 0.94 | 0.92 | 0.94 | 0.89 |
| Proximity | ||||||
| ln(Distance) | (0.77)**** | (0.21)* | (0.75)**** | (0.48)**** | (0.75)**** | (0.53)**** |
| Common Language | 0.23 | 0.46** | 0.55**** | -0.28 | 0.37*** | -0.22 |
| Metro–Colony | 0.69 | 2.54**** | 0.94**** | (omitted) | 0.73*** | (omitted) |
| Colony–Metro | -0.15 | 1.92**** | 0.51 | 1.28*** | 0.14 | 1.40*** |
| Economic Agreements | ||||||
| EIAs | -0.25 | 0.62** | (0.34)* | 0.64** | (0.38)* | 0.52 |
| EU | -0.36 | 0.16 | -0.33 | 0.18 | -0.099 | -0.28 |
| OECD | 0.2 | 2.36**** | 0.35 | 2.31**** | 0.75**** | 1.42*** |
| WTO | 1.15** | 0.039 | 0.86* | -0.062 | 0.87** | -0.078 |
| Taxation | ||||||
| DTTs | 0.45** | 0.72** | 0.41** | 0.78** | 0.52*** | 0.52 |
| No Withholding Tax | 0.54** | 1.22*** | 0.61*** | 1.51**** | 0.53** | 1.92**** |
| Constant | 14.26**** | 4.88*** | 13.98**** | 7.92**** | 10.10**** | 10.73**** |
Appendix A: Tax Havens
| Jurisdiction | OFC Score | Notes |
| Malta | 11 | EU* |
| Bahamas, The | 11 | * |
| Bermuda | 11 | ** |
| Cayman Islands | 11 | ** |
| Guernsey | 11 | ** |
| Jersey | 11 | ** |
| Panama | 11 | |
| Barbados | 10 | * |
| Cyprus | 10 | EU* |
| Isle of Man | 10 | ** |
| Liechtenstein | 10 | |
| Netherlands Antilles | 10 | NL |
| Vanuatu | 10 | * |
| Virgin Islands, British | 10 | ** |
| Singapore | 9 | * |
| Switzerland | 9 | OECD |
| Hong Kong | 9 | CN* |
| Gibraltar | 9 | EU** |
| St. Vincent and the Grenadines | 9 | * |
| Turks and Caicos | 9 | ** |
| Antigua and Barbuda | 8 | * |
| Cook Islands | 8 | NZ |
| Grenada | 8 | * |
| Ireland | 8 | OECD, EU |
| Luxembourg | 8 | OECD, EU |
| Monaco | 8 | |
| St. Kitts and Nevis | 8 | * |
| Belize | 8 | * |
| Nauru | 8 | |
| Andorra | 7 | |
| Anguilla | 7 | ** |
| Marshall Islands, Republic of | 7 | US |
| Mauritius | 7 | * |
| Bahrain, Kingdom of | 7 | * |
| Costa Rica | 7 | |
| Aruba | 6 | NL |
| Samoa | 6 | |
| Seychelles | 6 | * |
| St. Lucia | 6 | * |
| Dominica | 6 | * |
| Liberia | 6 | |
| Lebanon | 5 | |
| Niue | 5 | NZ |
| Macao | 4 | CN |
| Montserrat | 4 | ** |
| Malaysia | 4 | * |
| Maldives | 3 | * |
| United Kingdom | 3 | OECD, EU (>25% Agreement) |
| Brunei Darussalam | 2 | * (<25% Agreement) |
| Hungary | 2 | |
| Israel | 2 | * |
| Latvia | 2 | |
| Portugal | 2 | |
| United States | 2 | |
| Netherlands | 2 | |
| Philippines | 2 | |
| South Africa | 2 | |
| Tonga | 2 | |
| United Arab Emirates | 2 | * |
| Uruguay | 2 | |
| Belgium | 1 | |
| Germany | 1 | |
| France | 1 | |
| Iceland | 1 | |
| Italy | 1 |
Appendix B: Descriptive Statistics
| Variable | OFC ≥ 3 Real | OFC ≥ 3 OFC | OFC ≥ 6 Real | OFC ≥ 6 OFC | OFC ≥ 9 Real | OFC ≥ 9 OFC |
| Sample Size | 10,059 | 2,928 | 11,016 | 2,547 | 13,273 | 1,330 |
| FDI (Dependent Var.) Mean | 922.38 | 2209.65 | 1079.69 | 1970.77 | 1132.09 | 2426.02 |
| StDev | 10384.40 | 22288.73 | 12340.61 | 19834.05 | 13614.73 | 23041.20 |
| Min | 0 | 0 | 0 | 0 | 0 | 0 |
| Max | 563601.40 | 710852.30 | 563601.40 | 710852.30 | 626255.10 | 710852.30 |
| lnGDP (host) Mean | 4.83 | 4.83 | 4.86 | 4.84 | 4.68 | 4.68 |
| StDev | 1.99 | 1.99 | 1.98 | 1.99 | 2.07 | 2.07 |
| Min | 0.38 | 0.38 | 0.38 | 0.38 | -0.065 | -0.065 |
| Max | 9.58 | 9.58 | 9.58 | 9.58 | 9.58 | 9.58 |
| lnGDP/cap (host) Mean | 9.08 | 9.07 | 9.13 | 9.11 | 9.18 | 9.16 |
| StDev | 1.30 | 1.30 | 1.30 | 1.30 | 1.28 | 1.27 |
| Min | 6.17 | 6.17 | 6.17 | 6.17 | 6.17 | 6.17 |
| Max | 11.49 | 11.49 | 11.49 | 11.49 | 11.64 | 11.64 |
| ln(Distance) Mean | 8.17 | 8.32 | 8.18 | 8.32 | 8.22 | 8.20 |
| StDev | 0.85 | 0.80 | 0.84 | 0.82 |
