In recent years, a remarkable number of countries have entered into bilateral investment treaties (BITs) as an ostensible means of protecting and promoting inward foreign direct investment (FDI). Today, more than two thousand such treaties have been signed, involving countries of all levels of development and from all corners of the globe, and investors are increasingly using the treaties to sue host state governments before international tribunals for alleged violations of treaty provisions. One of the key empirical questions that has emerged from the BIT phenomenon is whether the treaties "work." The central premise of investment treaties is that states that agree to the disciplines and rigors of international investment law will enjoy benefits that offset the various costs. In exchange for giving up what might be called "policy space," or some measure of regulatory autonomy, host states expect, or hope, to receive increased flows of investment.

Scholars have devoted substantial energy to examining whether this so-called "grand bargain" has in fact been realized. Most of these studies follow a common research design. The number of BITs that a state has signed or ratified is counted up, with the resulting independent variable regressed against country-level FDI flow data - such as that included in the World Bank's World Development Indicators - usually using a pooled cross-sectional time series design. Unfortunately, the results of these various statistical exercises are inconsistent. Some studies show that BITs can have massive positive impacts on foreign investment; others show modest positive impacts; still others show no impact at all, or even a negative impact. While this inconsistency may in part derive from the fact that the various studies differ in data sources, model content, and precise statistical methodology, it is nonetheless discouraging that scholars have not yet been able to provide anything close to a definitive answer of whether BITs indeed achieve their central purpose: increased flows of investment.

In this Article, I approach the basic question of whether BITs work from a novel evidentiary perspective. As I have argued elsewhere, the ability to persuasively answer the "Do BITs work?" question using traditional large-sample statistical methods and existing data are probably limited. The "best" answer is likely to remain an inconsistent and uncertain one unless scholars begin to approach the study of the impact of BITs on foreign investment decisions by focusing on different data, different methods, and different levels of analysis. While these alternatives will suffer their own potentially severe methodological and conceptual problems, combining results across approaches may allow triangulation toward something approaching an accurate understanding of how much, if at all, investors care about BITs when deciding whether and where to invest.

I pursue three avenues of inquiry. First, I examine whether investment treaties appear to influence rankings of "political risk" provided by for-profit business consultants. If investment treaties are important elements in the foreign investment decision-making process because they protect against the risk of adverse political actions (like expropriation), it might be expected that companies whose line of business is to gauge such risks will incorporate the presence or absence of treaties into their evaluations. In other words, I use political risk rankings as a dependent variable, in substitution for FDI inflows, the more commonly used dependent variable.

Second, I present results from a small e-mail survey of providers of political risk investment insurance, in which I asked respondents to describe the extent to which investment treaties entered into their underwriting decisions. I posit that if BITs matter to investors because they successfully reduce political risk, we might expect insurers to take the treaties into account when deciding whether to issue investment insurance on particular terms.

Third, I present results from an original, mail-based survey of general counsel (GC) in large U.S.-based corporations, in which I ask respondents whether investment treaties influence their companies' decisions to invest. I propose that if investment treaties meaningfully impact FDI, that influence is likely to flow into a corporation's decision-making process through its GC's office, which acts as a repository of legal knowledge within the corporation, and which is typically tasked with advising on the legal implications of corporate decisions. If the GC's office has little knowledge or appreciation of BITs as risk-reducing devices, it is unlikely that non-lawyer corporate decision makers factor the treaties into their investment decisions.

The results of these three lines of inquiry provide evidence that BITs do not meaningfully influence FDI decisions. BITs are not strongly correlated with political risk rankings, and providers of political risk insurance only inconsistently take BITs into account when making underwriting decisions. Indeed, the majority of providers surveyed do not view BITs as relevant to their underwriting decisions. Finally, general counsel report relatively low corporate familiarity with, or appreciation of, BITs as risk-reducing devices.

These results by no means definitively prove that BITs never matter to investors when they decide whether and where to invest. Nor do they prove that BITs will not matter more to investors at some time in the future, as knowledge of BITs and confidence in the strength of their protections grows. What the study does suggest, however, is that grandiose claims about the empirically demonstrated ability of BITs to promote investment should be consumed with caution. BITs may influence certain investment decisions, but the results presented below suggest that they do not seem to influence many others. BITs have probably not been the primary cause - and perhaps not even a partial cause - of the massive increase in foreign investment to the developing world that began in the 1990s.