Why Rich Countries Win Investment Disputes: Taking Selection Seriously

Investor-state dispute settlement (ISDS) is a rapidly growing field of international law, due mainly to a proliferation of investment treaties among states that grant foreign investors the right to pursue binding arbitration against states for alleged violations of property rights. However, many have argued that poorer countries are systematically disadvantaged in arbitration since the costs of litigation and possible arbitrator bias in favor of wealthy capital exporters make it harder for these states to defeat investors’ claims. This article argues that while high-income governments tend to win about 20% more investment disputes than lowor middle-income governments, this disparity in win-rates can be explained by differences in the rates of early settlement between rich and poor governments. I find that developing country governments with high costs to litigation are about 22 percentage points more likely to settle a given dispute. This generates a selection effect among those disputes that are actually litigated. Investors with low-quality cases that could otherwise extract an acceptable settlement from a poorer respondent state are forced to litigate against wealthier governments, inducing a spurious positive correlation between respondent state income level and success rate. After adjusting for this selection process using a novel weighting approach, I find that wealthier respondent countries are no more likely to obtain a favorable ruling from an arbitration tribunal than poorer governments. Overall, while resource disparities among litigants do matter for outcomes in ISDS, observed differences in win-rates are not primarily due to the biases of arbitrators. Rather, wealthier governments have greater bargaining power when negotiating over settlements with investors, which affects the types of disputes for which arbitrators ultimately render awards. ∗Harvard University, Department of Government [email protected]. †The author thanks Matthew Blackwell, Marc Ratkovic, Dustin Tingley, Gary King, and participants at the 2017 New Faces in Political Methodology Conference at Penn State for helpful comments on previous versions of this paper.