Mining Causality: AI-Assisted Search for Instrumental Variables

TL;DR

This paper introduces a novel approach using large language models to automate the search for instrumental variables and other causal inference tools, significantly speeding up the process and exploring larger search spaces.

Contribution

It demonstrates how to leverage LLMs with specialized prompting strategies to identify valid IVs and control variables, enhancing causal inference methods.

Findings

LLMs can effectively identify potential IVs in economic case studies.

Multi-step role-playing prompts improve the relevance of LLM-generated variables.

The approach extends to control variables and running variables in various causal inference designs.

Abstract

The instrumental variables (IVs) method is a leading empirical strategy for causal inference. Finding IVs is a heuristic and creative process, and justifying its validity -- especially exclusion restrictions -- is largely rhetorical. We propose using large language models (LLMs) to search for new IVs through narratives and counterfactual reasoning, similar to how a human researcher would. The stark difference, however, is that LLMs can dramatically accelerate this process and explore an extremely large search space. We demonstrate how to construct prompts to search for potentially valid IVs. We contend that multi-step and role-playing prompting strategies are effective for simulating the endogenous decision-making processes of economic agents and for navigating language models through the realm of real-world scenarios, rather than anchoring them within the narrow realm of academic discourses on IVs. We apply our method to three well-known examples in economics: returns to schooling, supply and demand, and peer effects. We then extend our strategy to finding (i) control variables in regression and difference-in-differences and (ii) running variables in regression discontinuity designs.