Development status and decision-making in investment treaty arbitration

In this paper, I reproduce, replicate and extend two representative empirical studies that examine biases against developing states in investment treaty arbitration and arrive at contradictory conclusions. I find no evidence of biases against developing states using the same datasets and estimation strategies as the original studies. I find similar results when using alternative estimation strategies in the replication. However, when I change to a larger and more up-to-date dataset in the extension, I start to find consistent negative correlations between respondent states' development status and their likelihood of losing. The results remain robust after I take into account confounding factors that could be correlated with unobserved case strength, and use the Heckman probit selection model to correct for potential selection bias caused by settlement. The patterns indicate that there may exist biases against developing states in investment treaty arbitration. (C) 2019 Elsevier Inc. All rights reserved.