The ability to ensure compliance with investor-state arbitral awards is often regarded as one of the
strengths of the international investment regime. Yet, there have been few systematic studies of compli-
ance to assess the extent to which states have actually complied with adverse investor-state compensa-
tion awards. This paper presents a new dataset that enables empirical research on compliance with these
decisions; it is the first publicly available dataset to focus on what happens after awards are handed down,
and in this way complements other databases on international investment law. This paper explains the
data collection process (and its associated challenges), discusses the design choices made in selecting
inputs and variables, presents a descriptive overview of the data, and examines how variables can be
used in future research. Moreover, various cases are used as illustrations of the challenges of collecting
and coding data on post-award processes and we explore what missing data can tell us abou
It is often claimed that international investment arbitration is marked by a revolving door: individuals act sequentially and even simultaneously as arbitrator, legal counsel, expert witness, or tribunal secretary. If this claim is correct, it has implications for our understanding of which individuals possess power and influence within this community; and ethical debates over conflicts of interests and transparency concerning ‘double hatting’—when individuals simultaneously perform different roles across cases. In this article, we offer the first comprehensive empirical analysis of the individuals that make up the entire investment arbitration community. Drawing on our database of 1039 investment arbitration cases (including ICSID annulments) and the relationships between the 3910 known individuals that form this community, we offer the first use of social network analysis to describe the full investment arbitration community and address key sociological and normative questions in the literature. Our results partly contradict recent empirical scholarship as we identify a different configuration of central ‘power brokers’. Moreover, the normative concerns with double hatting are partly substantiated. A select but significant group of individuals score highly and continually on our double hatting index.