In international investor-state arbitration—as in domestic litigation—adverse inferences can play a critical role in promoting compliance with evidence production requests and in establishing truth. Nevertheless, arbitral tribunals have proved hesitant to resort to adverse inferences, particularly when asked to draw such inferences against sovereign states, and the use of adverse inferences remains beset by ambiguity and uncertainty. This article addresses the use of adverse inferences by arbitral tribunals in the context of arbitrations involving state parties and concerning the Middle East. A survey of these cases indicates that the lack of clarity surrounding the use of adverse inferences arises in part because tribunals do not always provide reasons for drawing (or refusing to draw) adverse inferences, have been inconsistent in applying and using them, and generally prefer direct evidence over inferential reasoning. A more methodical approach to requests for adverse inferences could mitigate these inconsistencies and ambiguities, promote greater transparency and predictability, and ensure a balanced and fair application of the device in investor-state disputes. We propose a framework for applying adverse inferences, drawing on their logical and legal underpinnings, the work of international arbitration scholars, and the jurisprudence of arbitral tribunals. By using this framework, tribunals can apply adverse inferences in a more coherent and consistent manner, thereby providing greater certainty to parties, promoting compliance with evidence production requests, and contributing to the uniform development of an international lex evidentia.