Record Details

Replication Data for: Essays on Causal Inference and the International Investment Regime

Harvard Dataverse (Africa Rice Center, Bioversity International, CCAFS, CIAT, IFPRI, IRRI and WorldFish)

View Archive Info
 
 
Field Value
 
Title Replication Data for: Essays on Causal Inference and the International Investment Regime
 
Identifier https://doi.org/10.7910/DVN/JMAQLD
 
Creator Strezhnev, Anton
 
Publisher Harvard Dataverse
 
Description The included files provide replication data for each chapter of the dissertation "Essays on Causal Inference and the International Investment Regime." The abstract is below:

In the last several decades, the growth of bilateral investment treaties (BITs) and the incorporation of investment provisions into regional trade agreements has led to the development of an international legal regime defining the rules states must follow with respect to investments made by foreign nationals. Among the most controversial elements of this regime are investor-state dispute settlement (ISDS) clauses, treaty provisions which allow foreign investors to initiate arbitration directly against states for alleged breaches of the treaty. ISDS proceedings are typically held outside of any national court and the arbitrators are selected from a small cadre of elite international legal professionals. Debates over the effectiveness of investment law and the legitimacy of ISDS raise a number of important empirical questions, such as whether investment treaties are actually effective in increasing foreign direct investment, whether developing countries are systematically disadvantaged by investor-state arbitration, and whether the arbitrators deciding these disputes are truly impartial or are influenced by extralegal factors.

The papers in this dissertation develop novel methodological tools in order to answer these three causal questions. The first paper evaluates the effect of bilateral investment treaties on foreign direct investment by United States firms using a differences-in-differences approach. It introduces a new differences-in-differences estimator for situations where outcomes are observed for more than two time periods. It shows how the conventional two-way fixed effects estimator yields misleading results when treatment effects persist over time and develops a new inverse propensity score weighting estimator that allows researchers to adjust for covariates without the need for restrictive assumptions on the time periods that can be affected by treatment. The second paper explores the problem of non-random attrition as applied to studies of legal disputes. It explains how analyses of dispute outcomes will be misleading when the treatment of interest also affects the propensity of a case to settle early and drop out of the data. Building on the "principal stratification'' framework, it develops a robust weighting method for adjusting for the bias due to attrition that is explained by observed covariates. It applies this approach to explain why richer countries appear to win more investor-state disputes and shows that most of the gap in observed win-rates can be explained by the fact that less wealthy countries are more likely to settle disputes early, which affects the distribution of cases that receive a final ruling. The third paper (co-authored with Sergio Puig) shows how experimental designs can supplement observational research for questions that cannot be answered from available observational data. Since investor-state arbitration mechanisms typically permit each party to appoint an arbitrator, questions have been raised about the impartiality of these party-appointed arbitrators. Unfortunately, because the vast majority of arbitral decisions are unanimous, evidence for bias is difficult to infer from case data alone. This paper presents results from a survey experiment of 257 arbitrators and arbitration experts in which respondents were asked to decide a hypothetical investor-state dispute. It finds that when arbitrators were told that they were appointed by one of the two parties to the dispute, they gave more favorable decisions to that party. These results suggest that professional arbitrators likely exhibit meaningful biases in their decision-making due to the party appointment mechanism.
 
Subject Social Sciences
causal inference, foreign direct investment, bilateral investment treaties, investor-state dispute settlement
 
Contributor Strezhnev, Anton