Predicting Outcomes in Investment Treaty Arbitration

Crafting appropriate dispute settlement processes is challenging for any conflict-management system, particularly for politically sensitive international economic law disputes. As the United States negotiates investment treaties with Asian and European countries, the terms of dispute settlement have become contentious. There is a vigorous debate about whether investment treaty arbitration (ITA) is an appropriate dispute settlement mechanism. While some sing the praises of ITA, others offer a spirited critique. Some critics claim that ITA is biased against states, while others suggest ITA is predictable but unfair due to factors like arbitrator identity or venue. Using data from 159 final cases derived from 272 publicly available ITA awards, this Article examines outcomes of ITA cases to explore those concerns. Key descriptive findings demonstrate that states reliably won a greater proportion of cases than investors; and for the subset of cases investors won, the mean award was US$45.6 million with mean investor success rate of 35%. State success rates were roughly similar to respondent-favorable or state-favorable results in whistleblowing, qui tam, and medical-malpractice litigation in U.S. courts. The Article then explores whether ITA outcomes varied depending upon investor identity, state identity, the presence of repeat-player counsel, arbitrator-related, or venue variables. Models using case-based variables always predicted outcomes whereas arbitrator-venue models did not. The results provide initial evidence that the most critical variables for predicting outcomes involved some form of investor identity and the experience of parties’ lawyers. For investor identity, the most robust predictor was whether investors were human beings, with cases brought by people exhibiting greater success than corporations; and when at least one named investor or corporate parent was ranked in the Financial Times 500, investors sometimes secured more favorable outcomes. Following Marc Galanter’s scholarship demonstrating that repeat-player lawyers are critical to litigation outcomes, attorney experience also affected ITA outcomes. Investors with experienced counsel were more likely to obtain a damage award against a state, whereas states retaining experienced counsel were only reliably associated with decreased levels of relative investor success. Although there was variation in outcomes, ultimately, the data did not support a conclusion that ITA was completely unpredictable; rather, the results called into question some critiques of ITA and did not prove that ITA is a wholly unacceptable form of dispute settlement. Instead, the results suggest the vital debate about ITA’s future would be well served by focusing on evidence-based insights and reliance on data rather than nonreplicable intuition