Lauge Skovgaard Poulsen
With the often costly and far-reaching implications when investors use Bilateral Investment Treaties (BITs) to adjudicate claims against developing countries, governments in the developing world need powerful arguments to justify that these treaties are in their national interest. One notable contribution to this debate has recently been put forth by political scientists Jennifer Tobin and Mark Busch, who set out to investigate the link between BITs and preferential trade agreements (PTAs). Using statistical techniques, they analyse annual data on pairs of developing and developed countries between 1960 and 2004 and conclude that BITs « raise the prospects of getting a North-South PTA with all the deeper and reciprocal obligations that these entail. »
If true, Tobin and Busch have revealed an exceptionally important finding that may add weight to the argument that BITs have a positive effect on investment flows. For while much of the empirical literature argues that BITs have not assisted developing countries in attracting foreign capital to any considerable extent, PTAs appear to have had a stronger impact on where, and how much, multinationals invest abroad. In other words, BITs may not impact investment flows extensively in and by themselves, but they might do so indirectly by increasing the likelihood of broader economic integration agreements.
This brief essay will argue, however, that in its current form Tobin and Busch’s claim is based on a flawed understanding of BITs’ practical implications, and their statistical analysis is almost entirely contradicted by the politics of investment treaty rule-making in most developed countries.
BITs as preferential agreements
Tobin and Busch’s argument is in two parts. First, the authors posit that Western multinationals lobby their governments to sign both BITs and PTAs as a way to protect their investments and reduce trade costs, respectively. But for developed country governments, BITs are ‘easier’ agreements to conclude than PTAs, as the latter are often met with opposition from their own protectionist constituents. So whereas wealthy states willingly pursue BITs with developing countries, they only pursue a PTA when it provides exceptionally large benefits. And this, they argue, is when the PTA can provide home state multinationals advantages vis-à-vis their foreign competitors. The twist to the argument, however, is that such advantages will be partly eroded if competing firms are also covered by BITs in the host state. This implies that support for a PTA will be greatest if the developing country has not already concluded too many BITs with other developed country governments. To increase the likelihood of a PTA, Tobin and Busch therefore argue that « a BIT is better than a lot ».
One of the key underlying assumptions is thereby that a BIT can give multinationals considerable preferential advantages when a developing country has not signed BITs with home states of competing firms. But as intuitive as this may sound, it misrepresents investment treaties’ practical implications. For the most part, BITs include broad and open-ended definitions of investments, and many expressively cover indirectly controlled investments. So if firms are seriously concerned that they do not have the same BIT-coverage as their competitors, they have the option of structuring their investment vehicles accordingly, for instance through holding companies in third countries. So while the extent of ‘treaty-shopping’ is not well understood, Barton Legum, the former head of NAFTA’s arbitration division, is probably correct in stating that:
« The reality that foreign capital is highly fungible and the breadth of the definitions of investor and investment thus combine to effectively transform the facially bilateral obligations of the BIT into an obligation that the host State must consider potentially applicable to all investors. »
The multi-layered corporate legal structure of modern multinationals thereby implies that while protections granted in BITs are perhaps preferential in theory, this is rarely the case in practise. Even a single BIT with the ‘traditional’ broad coverage of investors and investments should for all practical purposes be understood as involving obligations owed to every other state and investor. The notion that a multinational can retain future PTA preferences if developing countries do not sign BITs with other home states therefore appears somewhat out of sync with the potential coverage of the global BIT-network.
The politics of investment treaty rule-making
But while this questions the curve-linear relationship between BITs and PTAs suggested by Tobin and Busch-i.e. that a BIT is better than a lot-it doesn’t necessarily contradict their first proposition, namely that developed countries tend to couple their programs on BITs and PTAs. Indeed, there are pragmatic reasons for doing so. Modern PTAs often include investment protection chapters, so having already signed a BIT would make any subsequent PTA negotiations simpler. Also, a BIT negotiation could indicate whether the developing country partner is ‘mature’ enough to enter into the much more complex and time-consuming PTA negotiations. But while a country like Japan is in fact using BITs as ‘stepping stones’ to PTAs for these very reasons, does this really reflect a broader phenomenon? Should developing countries expect that by signing a BIT with a developed country, a subsequent PTA becomes more likely?
Let us start with the United States. While the United States Trade Representative (USTR) has occasionally mentioned that a BIT with the United States may lead to a PTA, it is doubtful that such a linkage has existed as a matter of policy. Veteran American BIT-negotiator, Kenneth Vandervelde, fails to mention any link between BIT and PTA negotiations in any of his extensive works on the US BIT program for instance. Also, no studies of US PTAs that this author is aware of can help sustain that BITs paved the way for American trade agreements in any substantial way. So when looking at the actual sequencing of US BITs and PTAs, it is no surprise that the pattern hardly provides strong evidence in favour of Tobin and Busch’s claim: out of the thirteen US PTAs with developing or transition economies (including NAFTA), only three were preceded by a BIT. These are the PTAs with Morocco, Panama, and CAFTA-DR (where BITs had been signed with El Salvador, Honduras, and Nicaragua), yet Tobin and Busch provide no evidence that the BITs in question were in any way connected with the subsequent trade negotiations.
However, even if American decision-makers should on occasion have thought of BITs as potential stepping stones to a PTA, one should recall that the US has signed less than fifty BITs. And while naturally important, these constitute a miniscule share of the global BIT-network. EU15 countries alone have signed more than a 1000. So when suggesting a general pattern in BIT policy making among developed countries, Europe is clearly the litmus test. Yet, here the suggested linkage between BITs and PTAs is even harder to follow. Recall that at least until the entry into force of the Lisbon treaty (on 1 December 2009); EU member states negotiated BITs individually while entering into PTAs as a group. But while 61 percent of the developed countries in Tobin and Busch’s analysis are EU member states, they fail to discuss what this implies for their theoretical model. Nor is it accounted for in their statistical analysis, which leads to a rather bold implicit claim: a (hypothetical) developing country with no other investment treaties in place can increase its predicted probability to enter into an EU PTA from 0.01 to 0.05 simply by signing a BIT with an EU member state. That the absence or presence of BITs have played such a crucial role for the choice of EU PTA partners would probably surprise most European policy-makers, and the authors quote no official reports, have made no interviews, or found any study of EU trade politics that can support it.
Also, if we go back to their second proposition-a few BITs increase the likelihood of a PTA, whereas too many do the opposite-the implication in the European context is that a developing country with only one BIT, say with Finland, is more likely to obtain an EU PTA than one that has signed BITs with most EU member states. To the extent that BITs have played any substantive role in European trade politics thus far-which is questionable-this is a rather counterintuitive suggestion to say the least, and there again appears to be no additional evidence to support it apart from their statistical model.
A BIT of econometrics is not enough
So while the work of Tobin and Busch is thought-provoking, innovative, and sophisticated-making it worthy to engage with-their results are difficult to reconcile with realities on the ground. Why might this be? Apart from the fact that they probably misrepresent BITs’ practical implications for multinationals, one reason could be their almost exclusive reliance on econometric techniques to ascertain government strategies in the international investment regime. While at times a legitimate empirical strategy if the right data is available, it is risky indeed when it is not. Qualitative methodologies are naturally fraught with their own limitations and risks of biases, but it is often prudent to check whether causal claims based on statistical correlations alone are in fact supported by alternative types of evidence. In this case they don’t seem to be. So before including the stepping stone argument among the justifications to sign (at least a few) BITs, governments in the developing world would be well-advised to wait for further studies that confirm the authors’ claims using alternative methodologies and sources of data. Given the remarks made here, this is likely to be a considerable challenge.
Author: Lauge Skovgaard Poulsen is a PhD-candidate at the London School of Economics. He can be contacted at [email protected]
 J. Tobin and M. Busch, « A BIT is Better Than a Lot: Bilateral Investment Treaties and Preferential Trade Agreements, » 62 World Politics 1 (2010), at 31.
 UNCTAD, The role of international investment agreements in attracting foreign direct investment to developing countries (Geneva: United Nations 2009), ch. 3.
 Tobin and Busch also argue that too many PTAs with other wealthy states make them less attractive as PTA partners. While the underlying logic for this part of their argument is indeed convincing, the present essay will purely focus on their conclusions concerning BITs.
 B. Legum, « Defining Investment and Investor: Who is Entitled to Claim?, » 22 Arbitration International 521 (2006), at 525.
 Ibid., at 524.
 This section is an extension of; L. S. Poulsen, « The importance of BITs for foreign direct investment and political risk insurance: Revisiting the evidence, » in: K. Sauvant (ed.) Yearbook on International Investment Law and Policy 2009/2010 (New York: Oxford University Press, forth.), footnotes 60-63 and accompanying text.
 Not-for-attribution telephone interview with former Japanese BIT-negotiator, July 2009.
 For instance, Tobin and Busch support their econometric findings with a USTR quote that its BIT-negotiations with Pakistan would lead to a subsequent PTA. But while this was also the expectation of several key Pakistani negotiators (various interviews, Lahore and Islamabad, Spring 2009), it is doubtful whether it was actually the US position. A USTR official notes: « Since Musharraf had vested so much political capital into the possibility of a PTA, we couldn’t openly state that we found it practically impossible to get through in Washington, so we played along. » (Comment at seminar on Pakistan’s BIT program, Johns Hopkins School of Advanced International Studies’ South Asia Program, May 11, 2009 [the author of this essay was the speaker]).
 E.g. K. Vandervelde, U.S. International Investment Agreements (Oxford: Oxford University Press, 2009), in particular 66-82,
 See e.g.; J. Schott (ed.) Free Trade Agreements: US Strategies and Priorities (Washington DC: Institute for International Economics, 2004).
 See again Vandervelde, op. cit.
 Tobin and Busch code an EU PTA as a series of bilateral PTAs with EU member states.
 Even for the variables Tobin and Busch are able to control for, there are not only serious endogeneity challenges but also a considerable problem of missing data, see Tobin and Busch, op. cit., at 18-23.
 See generally; R. Echandi, « What do Developing Countries Expect from the International Investment Regime?, » in: J. Alvarez and K. Sauvant (eds.), International Investment Regime: Expectations, Realities, and Options (Oxford: Oxford University Press, forth.).