Demand Estimation for Rationally Inattentive Consumers

Ruolei Zhu

Home/Archive/arXiv

arXiv · 2610.29600

Demand Estimation for Rationally Inattentive Consumers

Ruolei Zhu

View original source ↗Read source document ↗Export metadata ↓

Abstract

We develop a rational-inattention BLP (RI-BLP) model for demand estimation from aggregate market shares, allowing for endogenous prices and unobserved preference heterogeneity. We establish conditions for uniqueness of market-share inversion and construct a simulated RI-BLP GMM estimator with instrumental-variable moments and market-share constraints. Under regularity and simulation-accuracy conditions, the estimator is consistent and asymptotically normal as the number of markets grows, with the number of products fixed. Monte Carlo experiments show accurate recovery of mean utility coefficients, with less precise estimates of dispersion parameters. In a canned-tuna application, RI-BLP and logit yield similar average own-price elasticities. Unlike logit, the fitted RI-BLP model implies that a price increase can reduce demand even for similar products and shift demand toward the outside option.

Explore related subjects

econ.EM

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Ruolei Zhu. 2026-10-21. Demand Estimation for Rationally Inattentive Consumers. https://arxiv.org/abs/2610.29600

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

The Anatomy of Dependence in Multivariate Ordered Choice

When individuals make simultaneous decisions across multiple ordered dimensions, standard multivariate ordered choice models impose narrow bracketing, with each dimension decided as if the others did not exist. We develop a general rectangular structure model capturing broad bracketing while nesting the standard model. The framework introduces two layers of dependence, one through the threshold structure associated with decision rules and the other through the correlation of latent utilities. We provide microfoundations, prove identification, and discuss estimation. We showcase the model in applications to parental investment in children's education, cryptocurrency familiarity and expectations, and health insurance, where the latter separates moral hazard from adverse selection.

econ.EM↗

arXiv2026

Marital Sorting, Household Inequality and Selection

Using CPS data for 1976 to 2025, we explore how wage inequality has evolved for married couples with both spouses working full time full year, and its impact on household income inequality. We also investigate how marriage sorting patterns have changed over this period. To determine the factors driving income inequality we estimate a model explaining the joint distribution of couples' wages that accounts for spouses' employment decisions. This requires a multivariate distribution regression model that incorporates a double selection rule. We find that income inequality has increased for these households and increased wage sorting has exacerbated the inequality resulting from individual wage growth. We find that positive sorting partially reflects the correlation across unobservables influencing the wages of both members of the marriage. We decompose the changes in sorting patterns over the 50-year sample into structural, composition and selection effects and find that the increase in positive sorting primarily reflects the increased skill premia for both observed and unobserved characteristics.

econ.EM↗

arXiv2026

Evaluating A/B Testing Methodologies via Sample Splitting: Theory and Practice

We develop a theoretical framework for sample splitting in A/B testing environments, where data for each test are partitioned into two splits to measure methodological performance when the true impacts of tests are unobserved. We show that sample-split estimators are generally biased for full-sample performance but consistently estimate sample-split analogues of it. We derive their asymptotic distributions, construct valid confidence intervals, and characterize the bias-variance trade-offs underlying sample-split design choices. We validate our theoretical results through simulations and provide implementation guidance for A/B testing products seeking to evaluate new estimators and decision rules.

econ.EM↗