Causal Inference, Agency, and the Problem of Inherent Endogeneity
Year: 2025
Type: review
Abstract: Researchers often leverage exogenous variation in an independent variable in order to understand its effects, yet endogenous decision processes are central to many of the real-world phenomena we wish to understand. This review explores whether there are situations in which exogenous and endogenous variation in the same independent variable (e.g., a policy, treatment, or other action) may lead to different outcomes. I begin by laying out a conceptual framework for understanding these inherently endogenous causal processes, identifying three types of mechanisms through which they might arise and discussing their application to a range of empirical phenomena, such as institutional reform, community natural resource governance, and interstate conflict. I then suggest that learning about inherently endogenous causal processes requires researchers to place endogenous decision-making at the center of analysis rather than seeking to abstract away from it. I survey a range of methods (both positivist and nonpositivist) for doing so.
Author: Martin J. Williams
Institution: University of Michigan
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4416711716