Improving Reputation BIT by BIT: Bilateral Investment Treaties and Foreign Accountability

ABSTRACT

The literature on foreign direct investment (FDI) has paid an increasing interest to international institutions such as bilateral investment treaties (BITs), but whether BITs help attract FDI is an unsettled question. Building on the existing literature, this article argues that BITs can change investors’ perceptions and the corresponding investment they make because signing BITs signals the involvement of another powerful country that is able to compel the host government to comply. This implies that the effect of BITs is not constant across signatory countries: BITs are more effective when they are signed with rich and influential countries. Using monadic and dyadic FDI data, this article finds that BITs signed with powerful countries (defined as the top six largest economies) lead to an increase in FDI inflows (both from these signatory countries and from other countries). BITs signed with other countries, despite in a larger quantity, have little influence on FDI inflows.

KEYWORDS:

• Bilateral investment treaty
• foreign aid
• foreign direct investment
• foreign power
• international institutions
• OECD countries

Acknowledgments

A previous version of this article was presented at the 2012 Midwest Political Science Association annual meeting. The authors would like to thank Clint Peinhardt, Jennifer Tobin, the editors of International Interactions, and three anonymous reviewers for their helpful comments and suggestions on this article.

Supplemental data for this article can be accessed on the publisher’s website.

Notes

¹ See http://unctadstat.unctad.org/TableViewer/tableView.aspx?ReportId=88.

² One exception is the investment in the resource sector. Due to the extremely high profitability of natural resources, investors are willing to invest in resource-rich countries even though the level of political risks is usually high (Jensen and Johnston 2011).

³ See also Henisz (2002).

⁴ However, Aisbett (2009) points out that the positive relationship may be a correlation rather than causation. That is, countries are more likely to select into a treaty if they have more capital flowing between each other. Bergstrand and Egger (2013) also find that economically similar and geographically close countries are more likely to sign BITs.

⁵ Büthe and Milner (2009), Grosse and Trevino (2005), and Salacuse (1990) argue, for example, that BITs send a signal to investors that liberal economic policies will be undertaken beyond the details of the treaty.

⁶ For more on the signaling mechanism, see Abbott and Snidal (2000), Martin (2005), Morrow (1999), and Thompson (2006).

⁷ Jandhyala, Henisz, and Mansfield (2011) argue that BITs signed before the 1990s and after 2000 are primarily for the credible commitment purpose and that BITs signed between these two periods are mostly due to peer imitation.

⁸ For more on the logic of international institutions and credible commitment, see, for example, Dai (2005), Fearon (1997), Guzmán (1998), and Morrow (1999).

⁹ Politicians may have short-term interests in FDI for a variety of reasons. Electoral time horizons, for example, may be short enough for politicians to benefit from a BIT in the short run without bearing its long-run consequences (whether good or bad). For more on the relationship between time-horizons and BITs, see Ginsburg (2005).

¹⁰ Our following argument does not pit the commitment and signaling mechanisms against each other but rather uses both to understand the potential reputational effects of BITs.

¹¹ For a more extreme example of this logic see Biglaiser and DeRouen (2007), who argue that US military occupation of a host government may reduce investment risk for US investors. While signing a BIT may entail a less dramatic prospect of retaliation, the logic is similar.

¹² Even outside quantifiable losses, BITs may carry a normative deterrent. Elkins, Guzman and Simmons (2006:823–824) argue that “the home government has an interest in broader principles of good-faith treaty observance. Treatment that violates a BIT qualifies as a breach of the fundamental principle of international law: pacta sunt servanda (treaties are to be observed).”

¹³ Many of these examples are from disputes that occurred before the current period of international arbitration. Maurer argues that the pressure to defend investors has decreased as investors have become more able to dispute violations through international arbitration.

¹⁴ http://www.faces-adr.org/faces-thoughts/repsol-and-argentina-appoin.html.

¹⁵ Raszewski, Eliana, and Eduardo Thomson, YPF Posts Record Plunge as Argentina Spurns Repsol Claim, 18 April 2012, Bloomberg.com.

¹⁶ There are a variety of venues that arbitrate international investment disputes. Some venues are at international legal centers, while others are ad hoc. For each of these, there are likely even more disputes that are settled out of court. To date, of the 393 known international investment disputes (arbitrated at the ICSID, the SCC [Stockholm Chamber of Commerce], UNCITRAL [United Nations Commission on International Trade], ICC [International Criminal Court], among other venues), 304 have used a BIT between the home country and host country as the legal instrument in the arbitration.

¹⁷ Note that some BITs may remove the state further from arbitration, “depoliticizing” the dispute resolution process. But even in these cases, when BITs create new channels for investors to pursue host governments, the power of the home government will still loom, as a shadow, over the arbitration process and specifically over fulfillment of final payment of the compensation awarded by the court. Nevertheless, the extent to which BITs depoliticize remains unknown, and we leave this for further consideration.

¹⁸ To our knowledge, the only exceptions are Berger et al. (2011), Büthe and Milner (2009), and Yackee (2008). Yackee (2008) divides BITs into “strong” ones (that is, BITs with meaningful arbitration provisions) and “weak” ones, and Berger et al. (2011) consider BITs with binding dispute settlement and those without. Both articles suggest that BITs have little impact on FDI. Büthe and Milner (2009) weight BITs by two signatory countries’ GDP, which they call “power-weighted” BITs, and find a positive effect of both cumulative BITs and power-weighted BITs on FDI.

¹⁹ The authors thank an anonymous reviewer for this term.

²⁰ Many have described how lower levels of political risk incentivize foreign investment, including Collier and Patillo (2000); Graham, Johnston, and Kingsley (2015); Henisz (2000); Jensen (2008); Jensen and Johnston (2011); and Tomz (2007).

²¹ While some BIT studies use dyadic data to examine the effect of BITs on FDI (for example, Haftel [2010] and Hallward-Driemeier [2003]), we use both monadic and dyadic data. As Büthe and Milner (2009:189) point out, monadic analyses should be preferred to dyadic analyses when the theory predicts that BITs’ effect on FDI goes beyond the signatory country and also because the monadic data are usually of better quality.

²² A similar measure appears in numerous FDI studies—for example, Allee and Peinhardt (2011); Enders, Sachsida, and Sandler (2006); Lee (Forthcoming); and Li (2006), the first of which also examines the effect of BITs on FDI.

²³ There are some negative values in the FDI data, which cannot be dealt with by the log transformation, and we code them into zeros.

²⁴ Available at: http://investmentpolicyhub.unctad.org/IIA.

²⁵ The data on ODA are from the WDI database. While the result in Garriga and Phillips (2014) applies especially to postconflict scenarios and is conditional on whether the aid is geostrategically motivated, further research can investigate the degree to which aid, on its own, constitutes an alternative hypothesis.

²⁶ OECD FDI is the sum of FDI from 21 OECD countries that provide official development aid. The data on OECD FDI are from the OECD Web site.

²⁷ It is well known that the inclusion of both an LDV and country fixed-effects may cause Nickell bias (Nickell 1981), but this bias is pernicious only when T is small. As Beck and Katz (2011) point out, there is nothing harmful to have both an LDV and fixed-effects when T is greater than or equal to 20. Our first sample covers 37 years, and the second sample covers 21 years, so the bias will be very small, if any. The results also remain the same if we drop the LDV. The inclusion of lagged FDI to control for autocorrelation is also seen in other FDI studies (Jensen 2003; Powers and Choi 2012).

²⁸ Available at: http://www.oecd.org/statistics/.

²⁹ Those 21 countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Italy, Japan, Korea, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom, and the United States.

³⁰ The long-run effect can be calculated by , where is the coefficient for the independent variable and is the coefficient for the LDV (Enders 2004).