Problem
Research questions and friction points this paper is trying to address.
Granular Instrumental Variables
latent aggregate shocks
factor-loading space
equity demand elasticities
inelastic-markets hypothesis
Granular Instrumental Variables: Estimation and Inference
🤖 AI Summary
This study addresses the construction of valid granular instrumental variables (GIVs) in factor models subject to potential aggregate shocks. By characterizing the orthogonal complement of the factor loading space, it establishes a novel link between GIV validity and this orthogonal structure, yielding a feasible estimation procedure that neither requires knowledge of the factor loadings nor relies on a large cross-sectional dimension. The proposed method is accompanied by formal inference and specification testing procedures. By circumventing the conventional dependence on high-dimensional cross-sectional data, the approach substantially enhances the flexibility and applicability of instrumental variable methods. An empirical application to estimating stock market multipliers reveals pronounced heterogeneity in equity demand elasticities across investor sectors, providing refined evidence supporting the inelastic markets hypothesis.
📝 Abstract
We develop an estimation and inference framework for granular instrumental variables (GIVs) in models with latent aggregate shocks. Our key insight is that valid GIVs are characterized by the orthogonal complement of the factor-loading space. This characterization yields a feasible procedure for constructing GIVs when factor loadings are unknown and does not require a large cross-sectional dimension. We provide practical procedures for inference and specification testing, and apply the framework to estimate the aggregate equity market multiplier. Our empirical results reveal substantial heterogeneity in equity demand elasticities across investor sectors and may provide nuanced support for the inelastic-markets hypothesis.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Granular Instrumental Variables
Factor-loading space
Orthogonal complement
Latent aggregate shocks
Inference
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
